Mimo and Smart Antennas For 3g and 4g Wireless Systems May 2010 Final

1/135
TABLE OF CONTENTS
1. BACKGROUND AND SCOPE OF WHITEPAPER ........................................................................................... 14

2. BASE STATION ANTENNA EVOLUTION........................................................................................................ 19
2.1. EARLY TECHNOLOGY ........................................................................................................................................... 19
2.2. CURRENT TECHNOLOGY ....................................................................................................................................... 20
2.2.1. Extensive Usage ................................................................................................................................. 20
2.2.2. Moderate Usage ................................................................................................................................. 21
2.2.3. Advanced Antenna Technology .......................................................................................................... 22
2.2.4. MIMO Systems ................................................................................................................................... 23
2.2.5. Reconfigurable Beam Antenna ........................................................................................................... 24
2.2.6. Active Antenna .................................................................................................................................... 25
2.3. COMPARISON TABLE—ANTENNA TYPES SUPPLIED – ONE VENDOR ....................................................................... 25
2.4. ANTENNAS FOR MIMO ......................................................................................................................................... 26
3. ADAPTATIVE ARRAY (AA) BEAMFORMING ANTENNAS ............................................................................ 28
3.1. INTRODUCTION ..................................................................................................................................................... 28
3.1.1. Individual Antenna Column Characteristics ........................................................................................ 30
3.2. BEAMFORMING ..................................................................................................................................................... 35
3.2.1. Beamforming Antenna Calibration Requirements .............................................................................. 38
3.3. FIXED MULTI-BEAM ARRAY ANTENNAS ................................................................................................................. 41
4. RECONFIGURABLE BEAM ANTENNAS ........................................................................................................ 44
4.1. HOW RECONFIGURABLE BEAM ANTENNAS WORK ................................................................................................. 44
4.2. USE CASES OF RECONFIGURABLE BEAM ANTENNAS IN 3G NETWORKS ................................................................. 44
4.3. COMPARISON OF RET, 2D, AND 3D RECONFIGURABLE BEAM ANTENNAS .............................................................. 46
4.3.1. Measurement of Coverage, Interference, and Load Balancing with Reconfigurable Beam Antennas46
4.3.2. Network Optimization Versus Load Balancing.................................................................................... 47
4.3.3. Antenna Beam width Distribution ........................................................................................................ 49
4.4. RECONFIGURABLE BEAM ANTENNAS--CYCLICAL TRAFFIC PATTERN M ANAGEMENT................................................ 51
4.5. RECONFIGURABLE BEAM ANTENNA SUMMARY ...................................................................................................... 51
5. ACTIVE ANTENNA ARRAYS ........................................................................................................................... 53
6. ADDITIONAL SIMULATIONS AND COMPARISONS OF LTE TRANSMISSION SCHEMES / ANTENNA
CONFIGURATIONS .......................................................................................................................................... 58
6.1. COMPARABLE DOWNLINK SPECTRAL EFFICIENCY ................................................................................................. 58
6.2. AN ANALYSIS OF ANTENNA CONFIGURATIONS FOR 4X2 AND 4X4 MIMO................................................................ 59
2/135
7. DEPLOYMENT SCENARIOS ............................................................................................................................ 63
7.1. TYPICAL CELL SITE ARCHITECTURE ...................................................................................................................... 63
7.2. CURRENT DEPLOYMENTS ..................................................................................................................................... 64
7.2.1. Typical Tower Top Deployment .......................................................................................................... 65
7.2.2. Typical Mast Deployment.................................................................................................................... 67
7.2.3. Typical RooF TOP Deployment .......................................................................................................... 69
8. MISCELLANEOUS COMMERCIAL AND DEPLOYMENT ISSUES ................................................................. 72
8.1. CONSTRAINTS ON THE ANTENNA DEPLOYMENTS DUE TO COMMERCIAL CONSIDERATIONS ....................................... 72
8.2. ELECTRICAL AND MECHANICAL TILTING OF ANTENNAS .......................................................................................... 73
8.2.1. Effects of Mechanical Downtilt on Sector Antenna Horizontal Patterns ............................................. 73
8.2.2. Effects of Incorrect Antenna Installation ............................................................................................. 87
8.3. PASSIVE INTERMODULATION (PIM) SITE CONSIDERATIONS .................................................................................... 88
8.4. INDEPENDENT ANTENNA TILT OPTIMIZATION BY AIR INTERFACE ............................................................................. 89
8.5. REMOTE RADIO HEADS FOR MIMO ....................................................................................................................... 90
8.5.1. Remote Radio Heads and Transmitter-Receiver Duplexer Units ....................................................... 90
8.5.2. Remote Radio Head Advantages and Considerations ....................................................................... 92
8.5.3. Remote Radio Configurations ............................................................................................................. 92
8.6. FEEDER CABLE FOR REMOTE RADIO HEADS ......................................................................................................... 94
8.6.1. Hybrid Fiber/Copper Remote Fiber Feeders ...................................................................................... 97
8.6.2. Assemblies .......................................................................................................................................... 98
8.6.3. Lightning Protection ............................................................................................................................ 99
8.6.4. Accessories ......................................................................................................................................... 99
8.6.5. Advantages of Remote Fiber Feeders .............................................................................................. 100
8.7. CO-SITING OF MULTIPLE BASE STATIONS AND TECHNOLOGIES............................................................................ 101
8.7.1. Separate Feeders for Each System .................................................................................................. 101
8.7.2. Shared Feeders—Frequency Multiplexing ....................................................................................... 102
8.7.3. Shared Antennas .............................................................................................................................. 103
8.7.4. Shared Feeders and Antennas—Same Band Combining ................................................................ 104
8.7.5. Hybrid Combining ............................................................................................................................. 104
8.7.6. Low Loss Combiner—Multiplexer ..................................................................................................... 105
8.7.7. Single Carrier Power Amplifiers (SCPA) ........................................................................................... 105
8.7.8. Multi-Carrier Power Amplifier (MCPA) .............................................................................................. 106
8.7.9. Receiver Multicoupler ....................................................................................................................... 107
8.7.10. Integrated Devices—Same Band Combiner ..................................................................................... 107
3/135
8.7.11. Tower Mounted Amplifiers ................................................................................................................ 109
8.7.12. Benefits of Tower Mounted Amplifiers .............................................................................................. 109
8.7.13. Configurations ................................................................................................................................... 110
8.7.14. Enhanced Features—AISG .............................................................................................................. 110
8.8 MIMO CAPABLE DRIVE TEST EQUIPMENT ........................................................................................................... 112
8.8.1. Drive Test Overview.......................................................................................................................... 112
8.8.2. Scanning Receiver Measurements Versus User Equipment (UE) ................................................... 112
8.8.3. Impact of New Technologies on Drive Testing Scanners ................................................................. 114
8.9. INDOOR DISTRIBUTED ANTENNA SYSTEM—MIMO COVERAGE ............................................................................. 115
9. TERMINAL ANTENNA ISSUES ...................................................................................................................... 117
9.1. PROSPECTS AND CHARACTERISTICS OF MULTIPLE ANTENNAS IN TERMINALS ....................................................... 117
9.2. CURRENT STATUS .............................................................................................................................................. 121
10. SUMMARY AND CONCLUSIONS .................................................................................................................. 127
11. ACKNOWLEGEMENTS .................................................................................................................................. 128
12. ACRONYMS .................................................................................................................................................... 129
13. ENDNOTES ..................................................................................................................................................... 133
4/135
LIST OF FIGURES
Figure 1. Comparison of voice and data traffic trends showing the recent upsurge in wireless data traffic (Chetan
Sharma Consulting, March 2010) 11
Figure 2. Taxonomy of smart antenna processing algorithms in Release 8 of the LTE standard. 14
Figure 3. Antenna configurations (from endnote [6]) with the constraint of no more than four antenna cables per
sector for a total of 12 cables for a three-sector system. 16
Figure 4. Illustration of the terminology of base stations and cells. 17
Figure 5. Early base station antenna technology. 19
Figure 6. Current base station antenna technology -- extensive use. 20
Figure 7. Latest base station antenna trends -- moderate usage. 21
Figure 8. Advanced Antenna Technology. 22
Figure 9. MIMO Systems. 23
Figure 10. Reconfigurable beam antenna. 24
Figure 11. Active Antenna System (AAS). 25
Figure 12. Typical multi-column planar array architecture. 29
Figure 13. Beamforming antenna and fixed beam quad-port antenna. 30
Figure 14. Topology of a single column of radiating elements. 31
Figure 15. Connector ports for vertical and dual-polarized arrays. 32
Figure 16. Typical antenna patterns for a single column of elements. 33
Figure 17. Typical and relative antenna size for multi-column arrays at various frequencies. 34
Figure 18. Azimuth Pattern Beamforming. 35
Figure 19. Azimuth pattern beamforming with a scan angle of 20 degrees. 36
Figure 20. Null-forming and beam steering for optimizing sector performance. 37
Figure 21. Azimuth pattern beamforming comparisons of eight elements vs. four elements. 38
Figure 22. Beamforming degradation due to amplitude and phase errors. 39
Figure 23. Calibration network block diagram for a four-column beamforming array. 40
Figure 24. Single 65-degree antenna and twin fixed 33-degree beams. 41
Figure 25. Fixed twin-beam antenna architectures. 42
5/135
Figure 26. Comparison of the different degrees of freedom for the reconfigurable beam antenna family. 44
Figure 27. Life-cycle of a wireless network. 45
Figure 28. Compromise between coverage, interference, and load balancing effects in radio networks. 48
Figure 29. Simultaneous improvements achieved by ret, 2d reconfigurable beam, and 3d reconfigurable beam
antennas. 48
Figure 30. Distribution of the antenna beam width after the optimization. Case 1 is morning and Case 2 is afternoon.
50
Figure 31. Predictable changes in traffic distribution. 51
Figure 32. Active antenna array concept with a column array (left) and single antenna module (right). 53
Figure 33. Radio Integration Trend. 54
Figure 34. 900 MHz Active Antenna. 54
Figure 35. Typical active antenna architecture. 55
Figure 36. Electronic tilt applications. 56
Figure 37. Self-healing capacity. 56
Figure 38. Summary of downlink spectral efficiencies for various air interfaces and antenna schemes (3G Americas &
Rysavy Research, September 2009) 58
Figure 39. Downlink bit rate (left), downlink transmission rank probability (middle), and uplink bit rate (right) as a
function of the two dual bs antennas separation for the 4X4 and 1X4 antenna configuration in the DL and UL,
respectively. 60
Figure 40. Performance summary of different antenna configurations for DL and UL for networks in high or low load
conditions. 61
Figure 41. Results from full system simulation (left), single cell, single user simulation (middle) and field trial (right) for
downlink 4X4 antenna configuration. 61
Figure 42. Components of cell site, highlighting a single sector’s tower equipment. 63
Figure 43. Cell site under construction showing outdoor base station cabinets awaiting coaxial cable connection to
the tower mounted antennas. 64
Figure 44. Tower top “pinwheel” with dimensions. 66
Figure 45. Common monopole during installation of antennas. 67
Figure 46. Typical monopole antenna installation as seen from above and with remote radio heads mounted on the
monopole. 68
Figure 47. Mast with 11 operators sharing the same facility (including public safety and low power AM radio
broadcasting). 69
6/135
Figure 48. Typical rooftop installation with three sectors. the top scale spans 5 meters and the bottom 12 meters. 70
Figure 49. Antenna mount with remote radio head. 71
Figure 50. Angle below horizon for various tower heights. 74
Figure 51. How gain on the horizon changes with beam down tilt. 74
Figure 52. Differences between mechanically down tilted and electrically tilted antennas. 75
Figure 53. Differences in the horizontal pattern cuts. 75
Figure 54. Minimal horizontal pattern change using electrical down tilt only. 76
Figure 55. Combination of tilts. 77
Figure 56. Mathematical results plotted as a function of antenna’s elevation beam width. 78
Figure 57. Comparison mathematical results plotted as a function of antenna’s elevation beam width. 78
Figure 58. Variations in the K-factor which is 2.5 for 65-degree azimuth models ranges from 1.5 for 33-degree
Azimuth models to 3.3 for 90-degree azimuth models. 79
Figure 59. Horizontal Beam Squint. (Note the approximately 10-degree clockwise rotation of the beam compared to
the mechanical bore site.) 80
Figure 60. Beam peak versus mechanical tilt as a percentage of vertical beam width. 81
Figure 61. Varied mechanical tile with electrical tilt is fixed at zero. 82
Figure 62. Graphical representation of Sector Power Ratio (SPR). 83
Figure 63. SPR remains relatively constant for both v-pol and x-pol models as the electrical tilt is varied. 83
Figure 64. SRP degrades significantly for all types of antennas when mechanical tilt is varied which keeps electrical
tilt fixed at zero. 84
Figure 65. Front-to-back and Cross-pol ratio. 85
Figure 66. Demonstration of the huge amount of degradation that can happen to both f/b ratio and cpr when
mechanical down tilt is applied to an antenna having a large amount of electrical down tilt. 86
Figure 67. Azimuth Adjustment Tool. 88
Figure 68. Block diagram of a RRH/TRDU. 91
Figure 69. Examples of Remote Radio Heads with two transmit and receivers and with one transmitter and two
receivers. 91
Figure 70. Force on the cable vs. remote radio heads on a typical mast. this is assuming the 1-1/4” cables are
stacked two by six. the remote radio heads ranged from 12 kg to 20 kg and from 0.12 m 2 to 0.18 m2. 95
Figure 71. Photograph of the constrained access panel for running coaxial cable into a monopole. 96
7/135
Figure 72. Hybrid Fiber/Copper Cable. 98
Figure 73. Breakout and Termination Scheme. 99
Figure 74. Accessories. 100
Figure 75. Multiband sector with separate feeders. 102
Figure 76. Shared feeders using diplexers (crossband couplers). 102
Figure 77. Shared feeders and antennas using triplexers and broadband antennas, diplexers. 102
Figure 78. Diplexer. 103
Figure 79. Triplexer. 103
Figure 80. Hybrid Combining Cable Load Used to Ensure Low Passive Intermodulation. 104
Figure 81. Low Loss Combiner (LLC) with Integrated Duplexer. RX Distribution Provided from GSM BTS. 104
Figure 82. Multi-Carrier Power Amplifier (MCPA) with integrated RX distribution. 104
Figure 83. Single-Carrier Power Amplifier (SCPA) installed between radios and bts duplexer. 104
Figure 84. LLC combines narrow portion of TX band into broadband path. Includes duplexer for RX re-injection. 106
Figure 85. Filter multiplexer for downlink and uplink--quadriplexer. 106
Figure 86. SCPA module amplifies and combines two transmitters into one path. 106
Figure 87. MCPA for one sector with two duplexed inputs and two simplex inputs, three bricks installed. 107
Figure 88. MCPA for three sectors, two duplexed inputs and six simplex inputs per sector, one PA brick per sector,
and one hot standby brick. 107
Figure 89. Antenna sharing using TX/RX quadriplexer. 108
Figure 90. RX distribution to simplex BTS using duplexers and RXMC. 108
Figure 91. Antenna sharing using integrated SBC. 108
Figure 92. Dual RXMC with Eight Outputs per Cannel. 108
Figure 93. Integrated Same Band Combiner (SBC). 109
Figure 94. Single band TMA. 111
Figure 95. Twin single band TMA with AISG support. 111
Figure 96. TMA with integrated diplexer. 111
Figure 97. Dual band TMA with AISG support. 111
Figure 98. Diplexed dual band TMA with AISG support. 111
8/135
Figure 99. Dual diplexed dual band TMA with AISG. 111
Figure 100. Diplexed dual band TMA with pass-through and AISG. 111
Figure 101. Graphical Representation of the in-building coverage measured with: A) A D&F relay where 50 percent of
32
the link capacity is allocated for in-band backhauling, And B) An A&F optical DAS with an interleaved antenna
arrangement. 115
Figure 102. Graphs for MIMO link budget considerations: A) data rate enhancement MIMO vs. SISO as a function of
SINR, and B) typical indoor path loss graphics @ 2.6 GHz (WINNER II channel model). 116
Figure 103. Position of the handheld device for the two scenarios. 118
Figure 104. MIMO channel model. 120
Figure 105. Channel parameters. 120
Figure 106. Shannon upper limit capacity at 2.6 GHz and 780 MHz in an urban macro cell environment with two
transmit and two receive antennas. 121
Figure 107. EVDO data card clearly showing two-branch diversity. 122
Figure 108. Photograph of the inside of the Pantech UM150 EVDO USB dongle with 2 selectable antennas. 123
Figure 109. Example antennas from representative handsets, illustrating manufacturing and design approaches. 124
Figure 110. Internal view of the commercial CMU 300 modem with receive diversity. The two diversity antenna ports
are indicated. 125
Figure 111. Photo of the Inside of the Amazon Kindle showing two antennas for PCS band use. 126
9/135
LIST OF TABLES
Table 1. Antenna Types Supplied - One Vendor........................................................................................................... 25
Table 2. National cell site categories (all towers include the 3 following subcategories of towers). ............................. 65
Table 3. Number of Antennas per Sector. ..................................................................................................................... 65
Table 4. Example Wind Load for a Nine-Antenna (and Nine-cable) Tower with Ice Loading. ...................................... 96
Table 5. Link budget receive performance improvement due to two receive antennas over one receive antenna on a
handheld placed between a phantom head and hand. ................................................................................................ 118
Table 6. Link budget receive performance improvement due to two receive antennas over one receive antenna on a
handheld placed in front of user in a typical data usage fashion. ................................................................................ 119
Table 7. SCM parameters for 2X2 780 MHz antenna system. .................................................................................... 120
10/135
1. INTRODUCTION AND EXECUTIVE SUMMARY
According to Chetan Sharma’s analysis of the U.S. wireless market, wireless data traffic exceeded voice traffic in
15
2008 and grew by 193 percent in 2009 to over 400 petabytes (400 x 10 Bytes), about 4.4 times the information
1,2
content of the deep web.
Figure 1. Comparison of voice and data traffic trends showing the recent upsurge in wireless data traffic
(Chetan Sharma Consulting, March 2010)
With this rapid growth of wireless data traffic, now greatly exceeding voice traffic in many developed markets,
operators are anxious to quickly expand the capacity of their wireless networks. To address these demands for
increased capacity in a cost effective way, 3GPP standards have incorporated powerful techniques for using so-called
“Smart Antennas.” This paper outlines the considerable importance of various smart antenna schemes for improving
the capacity (and coverage) of the emerging generations of wireless networks.
Several insightful papers have described in detail the processing gains feasible with smart antennas; notably the 3G
3
Americas whitepaper entitled, “MIMO Transmission Schemes for LTE and HSPA Networks.” This paper concentrates
on the practical aspects of antennas and their deployment for 3G and 4G wireless systems.
Growth in data traffic is putting immense strain on the operator’s network. To increase capacity, operators have five
primary tools at their disposal:
1) Adding Cell Sites is an effective but expensive approach to adding capacity. In general adding
new real estate is time consuming and increasingly prohibitive. With median inter-site distances,
dropping from 5 km to 2 km and recently to less than 200 m in dense urban areas the operator has
less choice in selecting affordable property. Doubling the number of cell sites approximately
doubles the network capacity and the throughput per user (assuming the user density stays
2
constant), and greatly improves the peak user and the aggregate throughput per km .
11/135
2) Adding sectors such as changing from 3 sectors to 6 sectors is a useful way to approximate the
introduction of new cells. However, this does not quite double the capacity as the “petals” of 6
sector coverage do not interleave as well as 3 sector coverage, and the fractional overlap of 6
sectors is greater. This also challenges handoff processing when near highways. This is a
common approach in dense urban areas where rooftops are available. There is about a 70 percent
increase in capacity in moving from 3 to 6 sectors in an environment with low angle spread (where
the base station is located above the clutter).
3) Adding Carriers (or more accurately, bandwidth) directly adds to capacity. The LTE standard is
particularly adept at utilizing increased bandwidth without increasing control channel overheads. In
addition, in the U.S., the FCC permits increasing radiated power with the bandwidth in the PCS,
AWS, and lower 700 MHz bands providing improved penetration and coverage.
4) Improved air interface capabilities such as in evolving from Release 99 UMTS to Release 5
HSDPA that provided well over four times the aggregate downlink capacity. However, in moving
from Release 6 HSPA (1X2) to Release 7 (1X2) with 64QAM and 2X2 MIMO we see a more
modest 10 percent to 20 percent improvement in the aggregate throughput. As has been observed
before, improvements in air interface (while leaving everything else the same such as bandwidth
and antenna configuration) is seeing diminishing returns on improvements. Something more than
simply increasing modulation and coding rates is needed.
5) Smart antennas provide the next substantial increase in throughput. The peak data rates tend to
be proportional to the number of send and receive antennas, so 4X4 MIMO is capable of twice the
peak data rates as 2X2 MIMO systems. For another example, in upgrading from HSPA (1X2) to
LTE (2X2) a gain of 1.6x is seen (Rysavy, 2009).
By “smart antennas” we here refer to adaptive antennas such as those with electrical tilt, beam width and azimuth
control which can follow relatively slow-varying traffic patterns as well as so called intelligent antennas that can form
beams aimed at particular users or steer nulls to reduce interference and finally Multiple-Input Multiple Output (MIMO)
antenna schemes.
This paper focuses on the practical aspects of deploying smart antenna systems in existing Radio Access Systems
(RAS). A substantial and growing body of theoretical and field experience is able to provide reliable guidance on the
tradeoffs of various antenna configurations. Operational experience with WiMAX and TD-SCDMA systems that have
utilized smart antennas for several years provides practical experience applicable to the upgrade to HSPA and LTE
wireless systems. Many previous papers describe the theoretical capabilities of smart antennas and the mechanisms
that provide for their support in the standards. The reader is referred to surveys such as those in 1,4,5,6 (3G Americas
7
June 2009), and (BLTJ 2009 ) for details on simulations and measured performance of various configurations. A
summary of high-level results is also presented in Section 3.
This paper aims to address the practical tradeoffs of performance with the realistic constraints on the types of
antennas that can be realistically installed, cognizant of zoning, wind loading, size, weight and cabling challenges and
constraints from legacy terminals and other equipment. Constraints are, of course, present in both the base station
and the terminal side of the air interface, where MIMO technology promises useful gains if multiple antennas,
amplifiers, receivers and baseband processing resources can be made available in terminals.
Section 1 of this paper gives the background to the subject, defining terms and classifications of antenna schemes. In
particular, it discusses the LTE downlink transmission modes in relation to practical antenna configurations.
Section 2 concisely covers the evolution of the base station antenna. Legacy antennas are described, as are
antennas appropriate for multiband and multi-standard operation. Also described is a variety of advanced antenna
12/135
technologies used for receive and transmit diversity, beam steering and shaping, and MIMO. Active antennas are
introduced as well. Subsequent subsections provide more in-depth discussions of the various advanced antenna
technologies.
Section 3 discusses antennas used for MIMO and the suitability of dual-polarized antennas or spatially separated
vertically polarized antennas for a 2X2 application. Also discussed is 4X2 and 4X4 MIMO implementation using
horizontally separated dual-polarized antennas, and section 8.2 presents an analysis of the impact of different
antenna spacings on both the spatial multiplexing and the diversity performance.
Section 4 presents the planar array antenna typically used for adaptive array beamforming. The architecture,
application, calibration, and properties of these antennas is considered. The architecture implemented when planar
arrays are used to radiate multiple fixed beams is also explained.
Section 5 covers reconfigurable beam antennas, which extend the concept of variable elevation beam tilt to beam
control in the horizontal plane. The results of a radio planning simulation show benefits of optimizing both coverage
and interference, as well as the benefits of load balancing.
Section 6 introduces active base station antennas where the active RF functionality of the radio is integrated into the
antenna by distributing medium power transceivers directly behind each antenna element. This architecture and its
plusses and minuses are described.
Section 7 discusses spectral efficiencies of various air-interfaces and presents the results of simulation and field trials
on various MIMO antenna configurations.
Section 8 examines the typical deployment scenarios seen in current cellular networks.
Section 9 introduces various deployment issues including commercial restrictions on antenna and their use. The
accelerated horizontal pattern distortion, which is the result of mechanically downtilting an electrical tilted antenna, is
characterized. The topic of passive intermodulation (PIM) is explained as is its potential for interface in the RX path.
The remote radio head architecture and feature set for MIMO is described as is a RRH feeder cable solution. The
weight and windload advantages of a RRH installation as compared to a conventional coaxial feeder cable installation
is analyzed. Co-siting solutions for multiple base stations and technologies is treated using typical spectrum and
equipment scenarios. An overview of drive test equipment and the impact of multi-antenna technologies will have on
this equipment is discussed. Finally, two types of relay implementations are compared for extending MIMO coverage
indoors via a distributed antenna system.
Section 10 presents a study and its results relating realistic terminal antennas to the link budget and throughput that
can be expected for MIMO implementation in LTE. Practical examples of how terminal antennas are implemented in
hardware are presented.
13/135
2. BACKGROUND AND SCOPE OF WHITEPAPER
This white paper discusses downlink antenna techniques available in LTE Release 8. While additional techniques
such as Multi-User MIMO (MU-MIMO) are available for the uplink, and new techniques have been and will be
introduced in Release 9 and Release 10, they are beyond the scope of this paper.
Figure 2 below shows the taxonomy of antenna configurations supported in Release-8 of the LTE standard (as
described in 3GPP Technical Specification 36.300). The LTE Release 8 standard supports one, two or four base
station transmit antennas and two or four receive antennas in the User Equipment (UE), designated as: 1X2, 2X2,
4X4, where the first digit is the number of antennas per sector in the transmitter and the second number is the number
of antennas in the receiver. The cases where the base station transmits from a single antenna or a single dedicated
beam are shown in the left of the figure.
AC = Antenna Configuration
Decision Tree CL = Closed Loop
CQI = Channel Quality Indication
SISO, SIMO or MIMO= Multiple Input Multiple Output
2x2, 4x2 or 4x4
Linear Array Beam Forming / Space OL = Open Loop
Diversity Array
Division Multiple Access PMI = Pre-coding Matrix Indication
RI = Rank Indication
RS = Reference Symbol
SM = Spatial Multiplexing
Single Antenna or MIMO TD = Transmit Diversity
Array Array
Multi-User MIMO
Indicates
(AC5) Dynamic Adaptation
Adapts to OL TD
Mainly high Significant low &
vehicular speed pedestrian speed traffic
SISO or Planar Array
SIMO Beamforming Open Loop Closed Loop
Rank Indication (RI), Channel Quality Indication Rank Indication (RI), Channel Quality Indication
(CQI), no Precoding Matrix Indicaiton (PMI) (CQI), with Precoding Matrix Indicaiton (PMI)
Low SINR or low High SINR & Rich Low SINR or low High SINR & Rich
scattering Scattering scattering Scattering
Single Stream Multi-Stream

Multi-Stream Single Stream Multi-Stream
Multi-Stream
Rank 1 Rank
Rank 2-4
2-4 Rank 1 Rank
Rank 2-4
2-4
Port 0 Port 5 Closed Loop Closed

Closed Loop
Loop
Open
Open Loop
Loop
Dedicated RS Transmit Precoding Spatial
SpatialMult’xng*
Mult’xng
Common Spatial
Spatial Mult’xng
Mult’xng
(AC7) Diversity (AC6) (AC4)
(AC4)
Ref. Sig. (RS) (AC3)
(AC3)
Adapts to Common (AC2) Adapts
Adapts to
to CL
CL R=1
R=1
(AC1) RS & OL TD
Adapts
Adapts to
to OL
OL TD
TD Adapts to OL TD and
and OL
OL TD
TD
* Likely configuration for majority
of initial deployments
Figure 2. Taxonomy of smart antenna processing algorithms in Release 8 of the LTE standard.
Beyond the single antenna or beamforming array cases diagramed above in Figure 2, Release 8 of the LTE standard
supports MIMO antenna configurations. This includes Single-User (SU-MIMO) protocols using either Open Loop or
Closed-Loop modes as well as Transmit Diversity and MU-MIMO. In the Closed-Loop MIMO mode, the terminals
provide channel feedback to the eNodeB with Channel Quality Information through CQI, Rank Indications (RI) and
Precoder Matrix Indications (PMI), which supports the highest peak data rates, and is likely to be the most commonly
used scheme in early deployments. However, this scheme provides the best performance only when the channel
information is accurate and when there is a rich multipath environment. So Closed-Loop MIMO is appropriate in low
mobility environments such as with fixed terminals or those used at pedestrian speeds.
14/135
In case of high vehicular speeds, Open Loop MIMO may be used, but because the channel state information is not
timely, the PMI is not considered reliable and is not used. Note that in TDD networks, the channel is reciprocal and
thus the DL channel can be more accurately known based on the uplink transmissions from the terminal (the forward
link’s multipath channel signature is the same as the reverse link’s – both paths use the same frequency block), so
MIMO can potentially improve TDD networks under more conditions than in FDD networks.
One may visualize spatial multiplexing MIMO operation as subtracting the strongest received stream from the total
received signal so that the next strongest signal can be decoded and then the next strongest, somewhat like a multi-
user detection scheme. However, to solve these simultaneous equations for multiple unknowns, the MIMO algorithms
must have relatively large Signal to Interference plus Noise ratios (SINR), say 15 dB or better. With many users
active in a base station’s coverage area, and multiple base stations contributing interference to adjacent cells, the
SINR is often in the realm of a few dB. This is particularly true for reuse 1 systems, where only users very close to the
cell site experience SINRs high enough to benefit from spatial multiplexing SU-MIMO. Consequently, SU-MIMO works
to serve the single user (or few users) very well, and is primarily used to increase the peak data rates rather than the
aggregate data rate in a network operating at full capacity.
Angle of arrival beamforming schemes form beams which work well when the base station is well above the clutter
and when the angular spread of the angle of arrival is small, corresponding to users that are well localized in the field
of view of the sector, in rural areas, for example. To form a beam, however, one must have antenna elements spaced
less than a wavelength apart, while the spatial diversity required of MIMO requires either cross-polarized antenna
columns or else it requires that the columns be far apart. The farther apart, the more path diversity they will have.
This is often about 10 wavelengths (1.5 m or 5 ft. at 2 GHz).
These physical antennas are categorized below in Figure 3. The short lines correspond to individual antenna
elements, typically arranged in columns. Such columns are able to form the fan beam required to properly illuminate
a cell sector and which is a characteristic of base station antennas. The antenna elements in each column are
interconnected and share a common RF connector shown below the columns. These correspond to the individual RF
cables that connect the radios and their amplifiers. The configurations shown are restricted to no more than 4 cables
per sector, corresponding to the 4X4 limit in the Release 8 standard.
15/135
(A) (B) (C) (D) (E) (F) (G) (H) (I)
D D
Dual Bulter Matrix
Antenna Antenna
Algorithms Configuration Figure Description
AC1 1V (A) 1 Column with vertical polarization (V-Pol)
AC7 ULA – 2V (B) 2 Closely spaced V-pol columns
AC5, AC7 ULA-4V (C) 4 V-pol columns
AC2-4, DIV 1X (D) 1 Column with dual-slant-45 polarization (X-pol)

AC6
AC2-6 CLA-2X (E) 2 Closely spaced X-pol columns (Quad Port)
1 X-pol center column between two closely spaced x-

AC2-6 CLA-3X (F)
pol columns
AC2-7 BM-4X (aka (G) 4 X-pol columns with dual Butler Matrix
CLA-4X)
AC2-6 DIV-2X (H) 2 Widely spaced X-Polarized columns
AC2-6 TX-DIV (I) 2 Widely spaced V-pol columns
Figure 3. Antenna configurations (from endnote [6]) with the constraint of no more than four
antenna cables per sector for a total of 12 cables for a three-sector system.
The Butler matrix in antenna (G) is used to distribute properly phase and amplitude weighted contributions of the 4 RF
connectors to the eight columns to form four separate beams (two for each polarization, each half as narrow a
beamwidth as antenna (E)).
No one antenna configuration is applicable to all environments, for example in rural systems, where the eNodeB
antennas are located above the clutter, antennas that can form beams such as (C) and (G) are best. In urban
macrocellular antennas (E) (G) or (H) are expected to be best. Urban microcellular base stations that are embedded
in the clutter and the angle of arrival spread is large then antenna (H) is expected to be good at providing the greatest
path diversity, but as seen in Section 3 and in Figure 39, we see that the closer the columns of antenna (H) are, the
16/135
better the user data rates will be for downlink data, although different antenna spacings may be beneficial in different
environments or for different data types, etc.
Vocabulary
A word about vocabulary is in order. LTE is bringing with it a usage of the term “cell” that is more common among the
Europeans than that used in the Americas; namely, “cell” references the sector rather than the base station. Where it
has been common in the U.S. to refer to a “3-sectored cell,” it is becoming more appropriate not to use the term
“sector”, but to refer to the 3 “cells” associated with a single base station or Node-B. (One may also have other than
three cells associated with a single base station, four or six, for example.) In this document, we use the words “cell”
and “sector” interchangeably, and try to be clear in referring to base stations as a fixed site, such as a tower or
rooftop, where multiple antennas are located.
This is illustrated below in Figure 4 which shows how base stations are typically considered located at the vertices of
three cells (sectors). The interlaced “petals” of adjacent cells is such that the approximate beamwidth of the antennas
located at these base station sites are 60 degrees as illustrated in the bottom right part of the figure. These interlacing
petals work well for 3 sector base stations and show why antennas tend to be narrower in beamwidth than the 120
degrees one might otherwise assume from thinking of a base station located in the center of a cell.
Figure 4. Illustration of the terminology of base stations and cells.
The previous vocabulary will be increasingly strained as the industry deploys remote antennas further from the base
station, so we will refer to the area covered by common control signals (PCFICH, PDCCH, PHICH) as a cell.
We further refine our vocabulary by pointing out that a radome can contain multiple antennas, typically in columns
(though that is not the case with some active array antennas). These columns share a common RF connector with
the same RF signal. So a radome might consist of pair of cross polarized antennas one at -45 degrees, the other at
45 degrees polarization (from the vertical) and each with a distinct, RF connector that is fed with a distinct RF signal
having a separate Reference Signals (RS) while serving the same cell (sector). This is distinct from some references
17/135
that consider multiple antennas within a single radome to be part of one antenna. Both uses of the term antenna are
common on the cellular industry and this is common can lead to some confusion.
Some radomes serve multiple frequency bands with common RF connectors. These permit sharing of expensive
coaxial feeders and effectively have internal duplexers that split the different frequency bands and distribute the
signals to separate antenna columns. In this way, one can have a single coax and a single RF connector on the base
of a radome serving a multiple bands, with two or more antennas internal to the radome. The different bands might
be separately down-tilted and steered.
In the following section, we describe in detail the evolution and variety of antenna designs commercially available and
in development.
18/135
3. BASE STATION ANTENNA EVOLUTION
Base station antenna technology has progressed in response to industry requirements and trends. The key drivers
have been the continuing addition of cellular frequency bands, the integration of more functionality into single radome
housing, and antenna techniques that contribute additional capacity to cellular networks. The following figures
concisely describe the development of the base station antenna including advanced antenna technologies in use and
emerging today.
3.1. EARLY TECHNOLOGY
Early Technology
Base Station Antenna Evolution
OMNI DIRECTIONAL SECTORIZED LOG PERIODIC DIPOLE
1983 1989 1995
 First cell antenna  Increases capacity  Directive element

 Radiates equally  Azimuth beam widths 65º,  Focuses sector beams
 Low capacity 90º,105º  Improves handoff
 For RX spatial diversity  Sector-to-sector handoffs  Reduces interference
two antennas are  Coverage shaping
separated by 10λ provided by
 Mechanical tilt
 El beam width
Figure 5. Early base station antenna technology.
19/135
3.2. CURRENT TECHNOLOGY
3.2.1. EXTENSIVE USAGE
Current Technology—Extensive Usage

DUAL POLE & ELECTRICAL TILT VARIABLE TILT & RET MULTI BAND
1998 2003
2001
 Polarization diversity
 Internal phase shifter
replaces spatial diversity  DualBand combines low
 One dual, slant 45 controls variable beam tilt
band and high band arrays
polarization antenna  Beamtilt adjustment of the
into one radome
cell radius to optimize
replaces two vertical pol  Minimizes radome count,
antennas interference and handover
lease cost, wind loading, and
 Electrical tilt replaces  Motorizing the phase shifter
tower loading
for Remote Electrical Tilt
mechanical tilting  Each band has independent
 Electrical tilt superior due  RET avoids the cost of
RET for separate
to undistorted coverage as tower climbs when
optimization
optimizing the beam tilt
the beam tilts  TriBand versions are offered
Figure 6. Current base station antenna technology -- extensive use.
20/135
3.2.2. MODERATE USAGE
Current Technology—Moderate Usage

6 SECTOR MULTIBEAM ANTENNA CONCEALMENT &
INTEGRATION
-100 -90 -80

-110 -70
0
-120 -60
-130 -5 -50
-140 -10 -40

-15
-150 -30
-20
-160 -25 -20
-170 -30 -10

-35
180 -40 0
170 10
160 20
150 30
140 40
130 50
120 60
110 70
100 90 80
 Additional sectorization  3-sectors in one radome

increases capacity  Multi-beams from a single
array –butler matrix feed  Improved aesthetics and
 33º and 45º beam widths concealment
 Multiple antenna columns  Separate ports for each beam
 Independent tilt for each  TMAs can also be
creates the narrow beam integrated
 DualBands available but beam
not common  Reduces the number of
 RET antennas for a six sector site
from six to three
Figure 7. Latest base station antenna trends -- moderate usage.
21/135
3.2.3. ADVANCED ANTENNA TECHNOLOGY
3.2.3.1. ADAPTIVE ARRAY (AA)
Advanced Antenna Technology

ADAPTIVE ARRAY (AA)
• Planar Array – multiple columns of radiators • 4,6, & 8 column v-pol used for WiMax and
• External DSP controls the horizontal TD_SCDMA
antenna pattern • 4 column dual-pols are now also used – 8 ports
• A unique beam tracks each mobile • λ /2 spacing between columns is required as
• Pattern nulls mitigate interfering signals coherence is needed to form beams (correlation)
• Increased S/N generates capacity • Calibration is needed to eliminate phase and
improvements amplitude variation in feeder cables
• Often tower top electronics are used with AA
• AA is best suited for TDD and LOS applications
Figure 8. Advanced Antenna Technology.
22/135
3.2.4. MIMO SYSTEMS

MIMO SYSTEMS
2 x 2 SU- MIMO: Spatial Multiplexing
Same time and frequency resource Different data streams
• Multiple Input Multiple Output • Space Time Block Coding is a transmit diversity
• Capacity gains due to multiple mode used when S/N cannot support Spatial
antennas at both ends of the link multiplexing
• Multipath provides additional channel • Decorrelation between antennas and propagation
using DSP paths required for Spatial multiplexing
• LTE supports 1x2,2x2,4x2,4x4 • A Dual polarized BSA for 2x2 MIMO; two separated
• Spatial Multiplexing requires a multi- for 4x2 or 4x4 MIMO
path environment • Alternatively vertically polarized antennas can be
used with spatial separation
Figure 9. MIMO Systems.
23/135
3.2.5. RECONFIGURABLE BEAM ANTENNA

RECONFIGURABLE BEAM ANTENNA
Beam Panning
• Passive beam control

• 1-way: standard RET antenna
• 2-way: adds beam panning – horizontal beam
steering
• 3-way: adds beam fanning – horizontal beam
width adjustment
• Capacity improvements:
• - network optimization
Beam Fanning • - load balancing
• Dynamic coverage adjustments can be made
in response to changes in traffic distribution -
SON
Figure 10. Reconfigurable Beam Antenna.
24/135
3.2.6. ACTIVE ANTENNA

Active Antenna System (AAS)
M M
M M
M M
M M
M M
C
M M
M M
M M
M M
M M
CPRI/OBSAI
•Radio-embedded BSA • Input to BSA is via fiber and CPRI or OBSAI radio
•Functionality of a Remote Radio Head interface standard
integrated and distributed into the antenna • Electrical tilt beam control done digitally
•Same size as a standard passive BSA
• Electronic elevation beam control features that can
•Standard azimuth beam
provide capacity benefits:
•RX/TX Transceiver feeds each element directly
•Vertical sectorization
•Loss of a transceiver does not cause a
•Tilt by service or carrier
complete failure of the sector – graceful
•Separate TX-RX tilt
degradation with self-healing
Figure 11. Active Antenna System (AAS).
3.3. COMPARISON TABLE—ANTENNA TYPES SUPPLIED – ONE VENDOR
The following is a table comparing the distribution of antenna types recently supplied by one global vendor.
Table 1. Antenna Types Supplied -- One Vendor.
Antenna Type % Comments

Vertically polarized (V-pol) 10 80% of these are Omni; the rest are high gain
sector antennas for rural deployment
Dual polarized (X-pol) 90

Variable tilt 60-70
Fixed Tilt 30-40
25/135
RET 15-20 RET market share is increasing
Non-RET 80-85

Single band 60-70 40% of these are quad port
Multiband 30-40 Multiband market share is increasing
Antenna Type--Elevation Beamwidth 800/900 MHz

Degrees % Comment
16 45 Nominal height is 1.4M
4 0 N/A
Antenna Type--Elevation Beam Width 1800/1900/2100 MHz
Degrees % Comment
Antenna Type--Azimuth Beam Width
Azimuth
(Degrees) % Comment
90 15 Typically Rural
65 75 Urban/Suburban/Rural
33 and 45 20 Six Sector
3.4. ANTENNAS FOR MIMO
MIMO systems place the same requirements on the RF link as do the receive diversity systems that are in place for
current cellular networks, that is, there must be de-correlation between the channels received at the antenna. This
de-correlation is provided by space diversity when achieved by the separation of the antennas, or by the use of
polarization diversity when implemented by the use of orthogonal antenna elements.
Early cellular systems employed spatial diversity and typically used two vertically polarized antennas separated by a
distance of 10 wavelengths, or greater, at the frequency of operation. Most cellular providers have switched to
polarization diversity, utilizing dual polarized antennas, which have been shown to provide equivalent if not better
8
diversity gain as it does for spatial diversity. Dual polarized antennas have the added benefit of integrating two
antenna arrays into one radome housing while maintaining the same size.
Most antenna properties, and their associated specifications, influence the illuminated coverage of the cell site
topography and the link budget between the base station and handset. However, for dual-pol antennas, cross-polar
discrimination and port-to-port isolations can affect the diversity or MIMO performance of the system by introducing
correlation between the channels. Studies have shown that the standard specifications that meet the requirements for
26/135
9
effective receive diversity performance, also provide adequate de-correlation for effective MIMO system
10
performance.
In summary, a standard dual-pol antennas work well for 2X2 MIMO, as do two spatially separated dual-pol antenna
for 4X2 or 4X4 MIMO. A quad antenna, which packages two dual polarized arrays into one radome, provides
effective 4X2, or 4X4 MIMO performance in a compact width radome. Spatial separation of 1λ between dual-
10
polarized arrays is the norm for quad antennas. Studies presented in section 6.2 of this paper indicated that quad
antennas that have a spatial separation of less than 1λ can provide throughput gains for closed-loop, spatial
multiplexing, pre-coded beamforming (LTE transmission mode AC4) albeit at the possible expense of degraded
diversity performance and some compromise in antenna performance.
27/135
4. ADAPTATIVE ARRAY (AA) BEAMFORMING ANTENNAS
4.1. INTRODUCTION
An adaptive beamforming multi-column array antenna can be considered an advanced multiple antenna technique
that will provide an improvement in the overall communication link between the base station and mobile. The basic
architecture of an adaptive beam-forming antenna consists of multiple columns of radiating elements that are driven
by separate transceiver networks. Each transceiver path has the ability to adjust the amplitude and phase of the
transmit (downlink) and receive (uplink) signals.
Figure 12 below shows typical planar array architectures that are currently deployed in wireless networks. Three
different antenna array architectures are shown:
• Eight columns of vertically polarized radiating elements
• Four columns of vertically polarized radiating elements
• Four columns of dual-polarized radiating elements
There has been a significant trend to migrate towards dual-polarized arrays as the resulting physical size of the
antenna has been reduced by 50 percent as compared to a vertically polarized array.
It is important to note that the beamforming that we are referring to takes place in the azimuth plane of the beam.
This will be discussed further when we review various beamforming azimuth patterns.
As shown in Figure 12, the separation between the columns for a beamforming array is on the order of 0.5
wavelengths. It is not unusual to further increase the spacing to as much as 0.7 wavelengths, but there will be some
resulting degradation in the beamforming characteristics under various scanning conditions. By maintaining a fixed
column spacing of 0.5 wavelengths, grating lobes will be completely suppressed as the beam scans across the
sector.
28/135
Figure 12. Typical Multi-Column Planar Array Architecture.
Figure 13 below shows a comparison of another type of antenna topology that is often confused with a beamforming
type of antenna. Effective implementation of MIMO techniques requires that the signals on any ports not be highly
correlated. For the beamforming antenna shown below with 0.5 wavelength column separation, the signals between
adjacent columns would be highly correlated in low angle spread environments. The two-column antenna shown on
the right has a column separation on the order of 1.2 wavelengths. This type of antenna would be suitable for a 4X4
MIMO scheme, as the signals between any two sets of cross polarized ports would not be highly correlated even in
environments with relatively low angle spread. This two-column antenna could support a 2X2 MIMO scheme as well.
Conversely, the two-column antenna with 1.2 wavelength separation would be a poor choice for a beamforming
antenna.
29/135
Figure 13. Beamforming antenna and fixed beam quad-port antenna.
4.1.1. INDIVIDUAL ANTENNA COLUMN CHARACTERISTICS
Each of the individual columns of a multi-column beamforming array are comprised of radiating elements, a low-loss
power combining corporate feed network and an input port to provide access for a discrete transceiver. Figure 14
shows the topology of a single column of vertically polarized elements.
30/135
Figure 14. Topology of a single column of radiating elements.
Multi-column, adaptive, beamforming array antennas will therefore have a single output port for each corporate feed
network. For a vertically polarized antenna array, this will result in a single port for each column. For a dual-polarized
antenna array, each column will require two ports as a corporate feed network is dedicated for each polarization.
Figure 15 depicts a typical connector scheme for a four column vertically polarized antenna with four ports as
compared to a four column dual-polarized array that requires eight ports.
31/135
Figure 15. Connector ports for vertical and dual-polarized arrays.
The typical azimuth and elevation patterns of a single antenna column are shown below in Figure 16. The important
characteristics of the azimuth and elevation patterns are highlighted. Many of the characteristics are similar to those
of a traditional single column sector antenna used in non-beamforming systems.
The azimuth pattern is characterized to have a 65-degree half-power beam width and a good front-to-back ratio of 30
dB. It is also common to have a broader half power beam width of 90 degrees. However, there may be more
unwanted interaction between the antenna columns when the azimuth beam width is 90 degrees as the spacing
between the columns is only 0.5 wavelengths. While the 0.5 wavelength column spacing is preferred for good
beamforming characteristics, it is not unusual to increase the separation between the columns to 0.6 or 0.7
wavelengths. The increased separation reduces the mutual coupling between the adjacent columns and provides
more favorable and predictable individual column azimuth patterns.
For many of the beamforming planar array antennas deployed today, it is common to physically mount the radio
electronics (remote radio heads) directly to the back of the antenna structure. It is important to characterize the
azimuth antenna performance with the electronics present in the actual deployed condition. Certain parameters such
as the front-to-back ratio or front-to-side ratio may be influenced as compared to a stand-alone antenna.
32/135
Figure 16. Typical antenna patterns for a single column of elements.
To emphasize the importance of reducing interference from adjacent cell sites, first upper side lobe suppression is
typically specified to be a minimum of 17 dB to 18 dB. Fixed elevation beam down-tilt requirements will often be
specified at approximately 2 degrees below the horizon. Many engineers have now expressed interest in
implementing remote electrical down-tilt (RET) capability for individual columns to gain the advantages of proven RET
technology. It is also common to include some null filling for the main beam below the horizon. However, null-filling
values must be judiciously selected as excessive null filling may degrade the gain.
The elevation half-power beam width will typically be within the range of 6 to 7 degrees. The elevation beam width is
strongly related to the overall height (vertical dimension) of the individual columns. The size limitation of the antenna
therefore may result in a significant trade-off in elevation beam width. There is also the obvious frequency
dependence of the antenna size on frequency. Figure 17 shows the relative physical sizes of three different antennas
at the common operating frequency ranges of 850 MHz, 1900 MHz and 2500 MHz. Each of the antennas shown has
an approximate elevation beam width of 6.5 degrees to provide a relative comparison. It is apparent from figure 6 that
at the lower frequency range of 850 MHz, a multi-column beamforming antenna may become unacceptably large and
require some compromise in elevation beam width.
33/135
Figure 17. Typical and relative antenna size for multi-column arrays at various frequencies.
There are typically well-defined requirements for port-to-port isolation, return loss and passive inter-modulation
performance. The minimum isolation between any two columns of the antenna array is typically a well-defined
requirement that is derived from system level performance. Typical specifications for isolation between any two
columns may be on the order of 25 dB. For a vertically polarized antenna with 0.5 wavelength spacing between
adjacent arrays, the coupling can be significant and may have a relaxed requirement of 20 dB.
There is also an isolation requirement between the +45-degree and -45-degree polarizations of any single column for
a dual-polarized antenna array. Typically, this isolation requirement will also be on the order of 25 dB to 30 dB.
The return loss performance of the individual antenna columns must be carefully evaluated over the various scanning
conditions that will be used to synthesize patterns. As the amplitude and phase values of the individual columns are
varied, the return loss will also vary and may degrade to unacceptable levels. By judiciously designing the radiating
elements, these return loss degradation scenarios can be avoided.
34/135
4.2. BEAMFORMING
Figure 18 and Figure 19 below show some basic azimuth pattern beamforming scenarios for a four-column array.
Figure 18 depicts the resulting antenna pattern for the case of all four transceiver modules generating identical
signals. This is the case of a uniform amplitude and uniform phase distribution that will result with a main beam in the
direction of bore-sight.
Figure 18. Azimuth Pattern Beamforming.
Figure 19 shows the resulting antenna pattern for the case of the four-transceiver modules generating equal
amplitude signals but with a progressive phase variation of 68 degrees. The resulting pattern shows how the main
beam is effectively scanned 20 degrees in the azimuth plane as a result of the progressive phase of the signals.
35/135
Figure 19. Azimuth pattern beamforming with a scan angle of 20 degrees.
The patterns shown in Figure 18 and Figure 19 are also typical patterns that an antenna supplier would be required to
demonstrate compliance. As can be seen from Figure 19, side lobes can rise and beam widths can increase as
beams are scanned away from broadside. Therefore, compliance is typically specified using multiple scan angles.
By adjusting the beamforming vector value (in both amplitude and phase) of the individual signals presented to the
elements, a beam can be synthesized and scanned across the sector. By intelligently controlling the amplitude and
phase values, a variety of different pattern shapes can be generated.
In any case, the fundamental objective of an adaptive beamforming antenna is to provide significant enhancement in
the uplink and downlink communication streams:
• For the downlink case (when transmitting a signal), the objective of the beamforming array is to increase the
signal strength in the desired direction (while reducing interference to the undesired directions if possible).
• For the uplink case (when receiving a signal), the objective of the beamforming array is to improve the
receiver sensitivity in the direction of the desired signal (while reducing interference from the undesired
directions if possible).
Figure 20 below shows an example of a synthesized azimuth antenna pattern for an eight-column array. The pattern
shown has been created by applying judiciously selected amplitude and phase values to the individual
transmit/receive paths. The resulting pattern shows several main-lobes that are directed in desirable locations. The
pattern also shows how nulls are generated in the directions of interference.
The processing requirements for beamforming such patterns can be quite sophisticated. Typical requirements may
require amplitude and phase beam weight values to be updated on the order of 10 milliseconds. Complex algorithms
are used to support the pattern synthesis and may be further impacted by the complexity of the channel and the
number of subscribers. However-, useful information regarding the channel is available from the uplink reference
signals generated by the TD-LTE or WiMAX subscribers.
36/135
Figure 20. Null-forming and beam steering for optimizing sector performance.
The number of elements in the beamforming array will also affect the complexity of the beamforming patterns. In other
words, an antenna array consisting of eight elements will allow for a higher degree of pattern shaping as compared to
a four-element array. Obviously, an eight-element array would result in a larger overall antenna and would require
eight transceivers resulting in a significant increase in cost and complexity. Figure 21 below shows the difference
between a scanned beam for an eight-column array as compared to a four-column array. The eight-column array
provides a sharper higher resolution main beam. The eight-column array also demonstrates the ability to generate
more nulls. The eight-column array would require more complex algorithms, would have a larger overall form factor
and higher cost. These are obvious considerations that would need to be evaluated by the system designer.
37/135
Figure 21. Azimuth pattern beamforming comparisons of eight elements vs. four elements.
4.2.1. BEAMFORMING ANTENNA CALIBRATION REQUIREMENTS
Beamforming quality in the azimuth plane depends on the relative accuracy of the amplitude and phase values of
each transceiver. As with all multi-column beamforming antenna systems, there is some degree of error due to
undesirable variations between each transmit and receive path. Some type of calibration method must be
implemented to minimize the amplitude and phase errors between transceivers and antenna columns. Figure 22
below shows the comparison of a beam synthesized without any amplitude or phase errors as compared to a beam
synthesized with random amplitude and phase errors. For this particular example, random amplitude errors of +/-0.5
dB and random phase errors of +/-20 degrees were applied simultaneous to each transceiver of an eight-column
beamforming antenna array.
38/135
Figure 22. Beamforming degradation due to amplitude and phase errors.
As can be seen in the patterns in Figure 22, phase and amplitude errors can result in significant beamforming
degradation. The degraded patterns may result in undesirably high side-lobe levels, squinting of the main beam,
degradation in gain as well as losing the ability to accurately position nulls. Typical beamforming systems deployed
today require that amplitude variations be limited to +/- 0.5 dB, while phase variations are limited to no more than +/- 5
degrees.
Calibration networks for beamforming antennas can be implemented by integrating directional couplers on individual
antenna paths. The coupled outputs are then combined and connect to a dedicated calibration transceiver. A typical
calibration network block diagram for a four-column beamforming antenna array is shown below in Figure 23. By
selectively powering up individual transceivers, the amplitude and phase characterization of each antenna path can
be achieved.
39/135
Figure 23. Calibration network block diagram for a four-column beamforming array.
The calibration network shown in Figure 23 allows for periodic automated calibrations. Typically, these calibrations
would take place at periods of low traffic usage, but essentially can be performed at any time. It is important to note
that the power divider network and directional couplers must be carefully designed and calibrated such that they do
not contribute additional amplitude and phase errors. These types of networks require careful s-parameter
characterization at the factory level to ensure that adequate performance levels are achieved.
For beamforming antennas with additional columns, there will be an obvious increase in cost and complexity of the
calibration networks. Some lower cost calibration networks have eliminated the need for couplers, combiners and
dedicated calibration transceivers. These lower cost calibration networks exploit the strong mutual coupling of the
adjacent antenna columns to establish phase relationships between columns. A separate calibration transceiver is not
required, but instead, the main transceivers are used for calibration activity. While these types of calibration networks
are lower cost and lower complexity, they do require more complex algorithms for extracting the calibration data.
It is important to note that the amplitude and phase errors are proportional to the operating frequency. Beamforming
antenna systems operating at 2.5 GHz and 3.5 GHz will observe phase variation significantly higher than for a system
operating at 850 MHz. These increased errors are due to the natural tolerance variations of the transmission line
paths at higher frequencies.
40/135
4.3. FIXED MULTI-BEAM ARRAY ANTENNAS
Introduction
To address the need of increasing capacity in high-density macro-cell sites, fixed multi-beam antennas can provide an
effective solution using multiple fixed beams.
An example of fixed twin-beam antenna patterns is shown below in Figure 24. On the left, we see a typical 65-degree
sector azimuth (horizontal) antenna pattern. The patterns on the right show the paired asymmetrical azimuth beams
created by a twin beam antenna.
Figure 24. Single 65-degree antenna and twin fixed 33-degree beams.
For the patterns shown in Figure 24, a passive butler matrix type power-splitting network creates the fixed
beamforming. The power-splitting network can be designed to accommodate relatively broad bandwidths. Fixed twin-
beam antennas are commercially available today that operate from 1710 MHz to 2180 MHz, and with the beams
individually capable of variable electrical down tilt and RET. Figure 25 below shows the basic architecture of fixed
twin beam antennas for vertically polarized and dual-polarized four column antennas.
41/135
Figure 25. Fixed twin-beam antenna architectures.
The fixed twin-beam antenna has many advantages over the conventional approach of using two discrete antennas
for sector splitting:
• The fixed twin-beam antenna provides pre-set alignment of the main beams while providing the optimal
overlap. For the patterns shown above, the main beam pointing angles are perfectly fixed at -27 degrees and
+27 degrees.
• The pattern shaping is optimized for sector splitting as well as preserving the footprint of the original 65-
degree coverage. This optimization can be observed in the sector edge roll-off characteristics and azimuth
pattern side-lobe suppression.
42/135
• Three antenna enclosures provide six sectors of coverage resulting in reduced tower loading and favorable
aesthetic appearance. The reduced antenna count minimizes cap-ex and op-ex costs.
It is important to note that the azimuth patterns are fixed as compared to the multi-column beamforming antennas
previously discussed. However, the fixed-azimuth-beam antennas commercially available today incorporate remote
electrical down-tilt for adjusting of the elevation patterns. The electrical down-tilt is implemented such that each of the
antennas can be independently controlled for greater flexibility in network optimization.
43/135
5. RECONFIGURABLE BEAM ANTENNAS
5.1. HOW RECONFIGURABLE BEAM ANTENNAS WORK
Reconfigurable beam antennas extend the range of remote beam changes from a single dimension for elevation
beam steering (Remote Electrical Tilt, RET) to multiple dimensions. These antennas include the possibility to change
the boresight or azimuth direction (panning), as well as the beam width of the antenna (fanning) remotely (see Figure
26). Figure 26 includes additional terminology associated with the beam control functionality of these antennas (1D,
2D, 3D and RET, RAB, RAZ).
1D Reconfigurable beam (RET) 2D Reconfigurable beam (RET 3D Reconfigurable beam (RET,

[tilting] and Remote Antenna Azimuth RAS, and Remote Antenna
(RAS)) Beam Width (RAB))
[tilting + panning] [tilting + panning + fanning]
Figure 26. Comparison of the different degrees of freedom for the reconfigurable beam antenna family.
5.2. USE CASES OF RECONFIGURABLE BEAM ANTENNAS IN 3G NETWORKS
Reconfigurable beam antennas can be used for multiple purposes, including coverage, quality, and capacity
improvements. Several use cases are explored with a focus on load balancing capabilities.
To gain better insight into the real value of reconfigurable beam antenna technology, a radio planning analysis was
conducted. Data was used from a 129 site, 3G UMTS radio network in Brussels, Belgium. The study utilized a
commonly used radio network planning tool, including the planning data from the configuration data base that is geo
data, operational target values, and traffic distribution statistics.
All base station and link budget-related parameters (power levels, operating frequencies, mobile service
requirements, transmit and receive properties of the base station, as well as the mobile equipment, etc.) are based on
typical examples used by operators to plan and optimize their UMTS radio network.
44/135
In general, radio networks can be limited by coverage, interference, or radio resources. Networks in the first phase of
their life cycle (and deployment cycle) are typically limited by coverage or bad interference design.
In contrast, mature radio networks are typically limited by capacity and radio resources. A description of these is
depicted in Figure 27. The maximum traffic load is indicated by the red line at the 100 percent mark. Since different
base stations may use different radio equipment, and hence have different absolute capabilities for maximum traffic,
relative numbers are shown.
Phase 1 shows the network where traffic demand due to mobile users in the network is very low. Networks are
typically designed and deployed based on land usage information (clutter), as indicated in the color-coded clutter map
on the top left-hand side in Figure 27. The loading of individual sectors is also well below the maximum traffic that can
be handled in such cases. The network is typically limited by coverage and interference rather than capacity.
Phase 2 describes a more mature network where higher traffic demands occur. Once the number of subscribers and
the data rates for the individual services increase, the cell utilization increases. This is shown on the bottom right-
hand side of Figure 27. The traffic distribution in such cases is typically highly inhomogeneous within a network,
which is indicated by the top right-hand side, color-coded traffic density map. This results in many cells with a low
load, while other cells are overloaded, i.e., exceed the nominal limit of 100 percent maximum traffic load. An
increasing number of blocked and dropped calls are the consequence, leading to significant additional infrastructure
investments to handle the offered traffic.
Figure 27. Life-cycle of a wireless network.
Therefore, one of the key questions to be answered by the study is whether reconfigurable beam antennas can help
balance the load between different cells, leading to a combination of coverage, interference, and capacity
improvements.
45/135
5.3. COMPARISON OF RET, 2D, AND 3D RECONFIGURABLE BEAM ANTENNAS
To investigate the impact of Reconfigurable beam antennas on load balancing, we compared its performance with
conventional antenna technologies, including RET antennas using the network and traffic scenario as shown in Figure
27, phase 2 as the starting position.
This is a 3G UMTS network with 65-degree antennas. The mechanical tilts are used from the underlying 2G network.
The initial electrical tilts were set to 2 degrees. This is a typical deployment coverage approach for a 3G overlay on
an existing 2G network. Comparing three different reconfigurable beam antenna types and the range of parameters
used for this simulation:
• RET antenna: Fixed 65-degree horizontal pattern, elevation range of the remote tilt [0 degrees to 10
degrees], no azimuth changes;  Tilting only.
• 2D Reconfigurable beam antenna: Fixed 65-degree horizontal pattern, elevation range of 0 degrees to 10
degrees, remote azimuth changes in the range of [–30 degrees to +30 degrees];  Tilting + Panning.
• 3D Reconfigurable beam antenna: Elevation range of [0 degrees to 10 degrees], remote azimuth changes
in the range of [–30 degrees to +30 degrees], flexible antenna beam width in the range of [33 degrees to 120
degrees];  Tilting + Panning + Fanning.
5.3.1. MEASUREMENT OF COVERAGE, INTERFERENCE, AND LOAD BALANCING WITH

RECONFIGURABLE BEAM ANTENNAS
Reconfigurable beam antennas can provide multi-dimensional improvements in coverage, interference, and load
balancing.
• Coverage Measurement
o Pilot coverage: Sufficient signal coverage on the pilot (also called common control or broadcast
signal) is the basis for all wireless communications systems. The received signal strength of the pilot
tone at the mobile has to exceed the receiver sensitivity to ensure that a mobile can connect to the
network. Reconfigurable beam antennas can significantly help to increase the basic network
coverage.
o Service coverage: Different services may require different received signal levels so that a successful
service connection (a voice call or any data transmission) can be established between the mobile and
the network. These service coverage requirements depend upon the received signal level and the
received signal-to-interference ratio.
• Interference Measurements
o Improvements in C/I (Ec/Io): Interference is one of the key limiting factors in wireless systems. In
order to work properly, every data transmission requires a minimum signal to interference ratio. This
1
is usually expressed in terms of carrier to interference ratio (C/I), or Ec/Io .
1
Ec/Io = Average energy per chip over the total received power spectral density, including signal and interference in a CDMA system.
46/135
Improvements in the C/I (Ec/Io) result in better service coverage, services with higher data rate
throughputs, and improved service quality in general. Improving the C/I (Ec/Io) also means that lower
transmit power levels per connection are required for the same service, leading to significant
improvements in the overall network capacity.
o Reduction in the cell overlap or “Soft Handover”: Cell overlap in CDMA type systems is required for a
smooth handover for a user moving from one service cell to another, also called “Soft Handover.” As
mobile devices in Soft Handover mode receive the same, hence redundant, information from multiple
base station transmitters simultaneously, a large cell overlap leads to significant capacity reduction.
Cell overlap should therefore be kept as low as possible, but necessary as required by soft handover.
Reconfigurable beam antennas can be used to control both interference, as well as cell overlap in a highly
effective manner.
• Load Balancing Measurements
As shown in Figure 27 on the right-hand side, the distribution of the sector load is typically highly inhomogeneous
within the network. While some sectors experience very high traffic loads, the utilization of the majority of the sectors
is well below the maximum load.
o Measurement in sector utilization: In an ideal network, the maximum system capacity can be
achieved by equally balancing load between individual sectors. This means that load from highly
loaded sectors should be shifted to those that can easily handle additional traffic.
o Reduction of the number of overloaded sectors: The key benefit of load balancing is to increase
sector utilization, which reduces the number of overloaded sectors. This leads to significantly
improved overall network performance and increased capacity utilization without additional radio
channels, sectors, or cell sites there by reducing or delaying CAPEX.
It will be shown that reconfigurable beam antennas enable flexible, cost-effective balancing of individual sector loads
by allowing remote tilting, remote panning and remote fanning.
5.3.2. NETWORK OPTIMIZATION VERSUS LOAD BALANCING
Although various improvements can be achieved simultaneously, the performance improvements that are achievable
depend upon the available degrees of freedom of the reconfigurable beam system.
Coverage and interference improvements can be considered as classical optimization tasks. The tradeoff of capacity
11
and coverage or coverage and Interference levels has been considered for use in dynamic optimization. The results
for these tasks are shown in Figure 28 below. The results for these tasks are shown in Figure 29A and include:
• Coverage improvement
• Reduction in cell overlap
• Relative C/I (Ec/Io) improvements
47/135
Figure 28. Compromise between coverage, interference, and load balancing effects in radio networks.
The results of load balancing are shown in Figure 29B, and include:
• Increase in sector utilization
• Reduction of overloaded cells
Figure 29A Figure 29B
Figure 29. Simultaneous improvements achieved by RET, 2D reconfigurable beam, and 3D reconfigurable
beam antennas.
In Figure 29A, the relative improvements in coverage, C/I (Ec/Io), and reduction in the cell overlap are shown. In
Figure 29B, the relative increase in sector utilization and reduction of overloaded cells are shown as compared to the
initial scenario. While RET antennas are good for interference optimization and cell overlap reduction, their benefit for
coverage improvements and load balancing are limited as compared to 2D and 3D Reconfigurable beam antennas.
Starting from the initial settings, different antenna technologies have been used for both network optimization and load
balancing, simultaneously. Improvements quantified by this study of a particular network and its associated
environment and traffic pattern are summarized below:
48/135
• Improvements by 1D RET antennas
o Optimization gains: It can be observed that the key strength of RET antennas is the optimization of
C/I (Ec/Io) performance in a network (see Figure 29A). Cell overlap can be controlled similarly to the
interference in the network. However, in this case RET antennas have shown a limited ability to
improve the coverage of the network.
o Load balancing gains: As Figure 29B shows, RET antennas were not as effective for load balancing
as were panning and fanning antennas. In this case, modifying tilt was obviously not sufficient to
move traffic effectively from one sector to another. This resulted in a small relative improvement of
cell utilization. As a result, the number of overloaded cells was only reduced slightly, see Figure 29B.
• Improvements with 2D Reconfigurable beam Antennas
o Optimization gains: In cases where 2D reconfigurable beam antennas were used, all optimization
objectives were improved when compared to the RET situation. C/I (Ec/Io) and cell overlap
performance have been improved, and the relative coverage improvement has more than doubled
compared to the RET case. See Figure 29A.
o Load balancing gains: Since 2D reconfigurable beam antennas have the additional degree of
freedom for remote changes of the boresight direction (panning), it is shown that the offered traffic
can be shifted more effectively between the individual cells. This resulted in better load balancing,
and higher cell utilization, hence the ability to reduce the number of overloaded cells significantly, as
shown in Figure 29B.
• Improvements with 3D Reconfigurable beam Antennas
o Optimization gains: In addition to 2D reconfigurable beam antenna cases, the beam width of the
antenna can be modified remotely (fanning). All of the performance indicators related to optimization
were significantly improved compared to the 2D reconfigurable beam case. C/I and coverage were
further improved. In this analysis, compared to RET antennas, C/I improvements were almost
doubled.
o Load balancing gains: Since 3D reconfigurable beam antennas have the highest flexibility to balance
the load between sectors, i.e., to shift traffic among the cells, the study indicated that the increase in
the average cell utilization was boosted. This lead to a dramatic reduction in the number of
overloaded cells, as shown in Figure 29B.
5.3.3. ANTENNA BEAM WIDTH DISTRIBUTION
Most existing wireless networks use a fixed choice for antenna beam width, that is, the 65-degree antenna which is
the most widely used antenna pattern worldwide. In contrast, this analysis shows that significant gains can be
achieved with the flexible beam width provided by the 3D reconfigurable beam antenna.
To give insight into the most appropriate antenna beam width, Figure 30 shows the statistical distribution of the
antenna beam width using the 3D reconfigurable beam antenna from the cases in the previous sections. Two beam
width distributions are shown for different traffic scenarios. The traffic scenarios are taken from different network
cases in the morning and late afternoon, respectively.
From the first look, it can be seen that the “65-degree pattern fits all” maxim is no longer valid. It has already been
shown from the results of the 3D reconfigurable beam antenna in Figure 30 that significant improvements of
49/135
coverage, C/I (Ec/Io), as well as a large reduction in the number of overloaded cells, were achieved by the use of a
flexible beam width.
A narrow antenna beam width results in a higher antenna gain in the main lobe direction, as compared to a wider
antenna beam width with a significantly lower main lobe gain. This is one reason why a mixture of different antenna
beam widths provides the best performance.
18
16
14
Occurrence rate [%]
12
10
0
35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120
Antenna beamwidth [°]
Traffic Case 1 Traffic Case 2
Figure 30. Distribution of the antenna beam width after the optimization. case 1 is morning and case 2 is
afternoon.
Distribution of the antenna beam width shown in Figure 30 shows some clusters around 45 degrees and at the upper
end above 100 degrees. The distribution depends on various input parameters, such as the scenario itself, the traffic
distribution, as well as the desired focus on the optimization objective, as indicated in Figure 28. Compromise
between coverage, interference, and load balancing effects in radio networks.
It is possible that wide beam width pattern will cause higher sector-to-sector interference. The results in Figure 29A
show a different picture: The actual sector-to-sector interference was not significantly reduced by the use of variable
beam width. This can be attributed to the fact that it is not just the beam width, but the combination of beam width
and azimuth changes that provides substantial improvements.
50/135
5.4. RECONFIGURABLE BEAM ANTENNAS--CYCLICAL TRAFFIC PATTERN MANAGEMENT
Measurements in wireless networks show a clear daily cyclical behavior. An example for the traffic pattern of a single
sector, over a 24-hour, 7-days a week period is shown in Figure 31. It can be observed that the traffic pattern repeats
over time.
Predictable Changes In Traffic Distribution
Sector A Sector B
Relative Traffic Per Hour
kbit/s
Mon Tue Wed Thu Fri Sat Sun
• Network traffic is neither static nor chaotic, but dynamic and predictable.
• Typical traffic hot spots for business hours, residential hours, rush hour.
Figure 31. Predictable changes in traffic distribution.
Keeping in mind that the overall traffic will grow over time, in general, the total traffic will be a combination of a cyclical
pattern combined with an increasing average value.
By using smart RF technologies, such as SON, with reconfigurable beam antennas, wireless networks can remotely
adapt to these continuously changing traffic situations. This adaptive, agile beam control technology has the potential
for a positive impact on both CAPEX and OPEX for networks operators.
5.5. RECONFIGURABLE BEAM ANTENNA SUMMARY
Reconfigurable beam antennas have been shown to be a benefit for flexible wireless network management.
The following conclusions can be drawn from the specific network analysis case presented.
• RET antennas had a significant impact on the improvement of C/I and Ec/Io in CDMA based 3G radio
networks.
51/135
• 2D reconfigurable beam antennas showed significant improvements in all performance indicators when
compared to the Brussels network optimized with RET antennas. This includes coverage and C/I (Ec/Io)
improvements, as well as higher cell utilization.
• Higher cell utilization allowed 2D reconfigurable beam antennas to shift traffic from overloaded sectors to
those that can carry extra load, in this case, reducing the number of overloaded cells by up to 50 percent.
• 3D reconfigurable beam antennas generated the highest performance gains of all the reconfigurable beam
antenna types. This was the case for coverage and C/I (Ec/Io) objectives, as well as for cell utilization and
the balance of cell loads.
• In the case of load balancing capability and improvements, for this typical example, 3D reconfigurable beam
antennas outperformed RET antennas by more than a factor of five.
• The optimal distribution of the antenna beam width indicated that the singular use of a 65-degree pattern
maybe far from ideal for wireless networks.
• In this case, reconfigurable beam antennas reduced the number of overloaded cells by approximately 80
percent. Higher utilization of existing sites, enabled by reconfigurable beam antennas, could reduce or delay
future CAPEX and OPEX expenses for an operator.
• 3D reconfigurable beam antennas allow highly increased flexibility. With cyclically changing traffic patterns, a
scheduled network adaptation, has the potential to impact the quality and cost and cost of cellular network
operation. Reconfigurable beam antennas, in combination with the network management systems are well
suited for this application.
12,13
References of interest on reconfigurable beam antennas can be found in the endnotes.
52/135
6. ACTIVE ANTENNA ARRAYS
A most general approach to DSP controlled smart antennas that permits both the typical horizontal beamforming but
also vertical beamforming, is made possible by active antenna arrays as shown in Figure 32. This scheme permits
amplitude and phase variation to be applied to vertically stacked antenna elements, such as the 16 elements shown
below. In contrast to standard passive base station antennas, where a coaxial cable distribution network divides and
feeds power to each element, individual transceivers, composed of radio modems, amplifiers, and filters are located
directly next to the radiating antenna elements.
Figure 32. Active antenna array concept with a column array (left) and single antenna module (right).
This technology has the added benefit of eliminating power losses in the RF feeder cables, much like Remote Radio
Heads. With the radios integrated directly into the radome housing, and with replacement of a small number of large
amplifiers with many small amplifiers, the heat is spread over the larger antenna structure as opposed to the smaller
RRH or amplifier shelf, permitting larger total RF transmit powers without the use of fans or other active cooling.
Alternatively, the use of many lower power amplifiers, operating at cooler temperatures, can increase the reliability of
the radio system.
By integrating the remote radio head functionality into the antenna, the aesthetics of the site can be improved,
windload reduced, and potentially, some leasing costs avoided. See Figure 33.
53/135
Figure 33. Radio Integration Trend.
Research and development activities on active antenna systems for cellular applications have been conducted by
14,15 16
various organizations over the last decade, and recently field trials have been completed. Figure 34 presents an
active antenna used in a 900 MHz field trial.
Figure 34. 900 MHz Active Antenna.
54/135
One architecture for an active antenna is depicted in Figure 35. In this architecture the distributed transceivers, or
micro-radios, are fed digitally from a central digital processing controller and are composed primarily of three custom
IC building blocks and a duplexer.
Figure 35. Typical active antenna architecture.
Another critical benefit of an AAS is the unique ability to electronically tilt elevation beams by having independent
base band control of the phase and amplitude on each element. This supports multi-mode systems where different
carriers in the same frequency band, with different air interfaces, may require different orientations. For example,
legacy GSM/GPRS/EDGE or UMTS carriers may provide adequate coverage, but when a newly introduced LTE
carrier is used for hot spot coverage or is used to span a more sparsely deployed area, LTE may be down-tilted
differently than the legacy carriers. With a two-column active array, the LTE carrier may also be directed toward
different azimuths directions than are the legacy carriers. The electronic tilt capability also allows for the separate
beam tilting and optimization of the TX and RX paths and of the vertical sectorization of a cell (see Figure 36). Vertical
17
sectorization of an LTE sector has been shown in simulation to provide significant capacity benefits.
55/135
Figure 36. Electronic tilt applications.
The distributed and redundant architecture of the AAS, where each antenna element has its own transceiver, provides
reliability benefits as the failure of one micro-radio does not cause the catastrophic failure of the sector. Of course,
some loss of both EIRP and received signal level is experienced, but this may be tolerable for a period of time while
the AAS awaits repair. Since the system is intelligent, and can sense a transceiver failure, the amplitude and phases
on the remaining elements can be adjusted to compensate for the elevation beam distortion and the reduction of EIRP
on the horizon (see Figure 37).
Figure 37. Self-healing capacity.
These newly proposed active array antennas are only now in trials and have not yet proven their reliability, the
acceptability of weight and their ability to economically linearize amplifiers sufficiently to make multi-channel emission
specifications without channel-wide filters. The power consumption of many low-power radio amplifier modules versus
56/135
a small number of large power amplifiers is also not yet clearly proven. However, similar active antenna arrays have
18
been used in military RADARs for many decades and have advantages for versatility and performance.
57/135
7. ADDITIONAL SIMULATIONS AND COMPARISONS OF LTE TRANSMISSION SCHEMES / ANTENNA
CONFIGURATIONS
7.1. COMPARABLE DOWNLINK SPECTRAL EFFICIENCY
Previous measurements, simulations and estimations of the relative spectral efficiency of various air interface
technologies and antenna schemes are summarized in Figure 38.
Figure 38. Summary of downlink spectral efficiencies for various air interfaces and antenna
schemes (3G Americas & Rysavy Research, September 2009)
These values are from a joint analysis by 3G Americas’ members based upon 5+5 MHz for UMTS-HSPA/LTE and
CDMA2000, and 10 MHz DL/UL=29:18 TDD for WiMAX, and assuming a mix of mobile and stationary users.
This figure clearly shows the relative performance of HSPA vis-à-vis LTE where HSPA with type 3 terminals
implementing Mobile Receive Diversity (MRxD) effectively double HSPA’s spectral efficiency from 0.5 bps/Hz/sector
to 0.9 bps/Hz/sector while LTE with 2X2 MIMO provides 1.5 bps/Hz/sector. The 2X2 MIMO gives HSPA a 20 percent
increase over MRxD, but 64 QAM capable terminals and Successive Interference Cancellation (SIC) can raise HSPA
efficiency to 1.3 bps/Hz/sector. Further improvements can be obtained in Release 9 HSPA with dual-carrier operation
with MIMO. Note that the upgrade to 64QAM can be implemented with a software upgrade in most base stations while
58/135
MIMO requires a change in antennas, though the contribution of 64QAM modulation is slight, due to the rarity with
which it can be expected to be available in typical cellular network operation.
It is important to note, that the gain of 2X2 MIMO in the case of HSPA+ assumes that all terminals have two receive
antennas. In a deployment where there are legacy terminals without 2 receiver and MIMO capability, the multi-stream
transmissions from the base stations transmitting SU-MIMO signals will contribute multi-path interference to older
terminals and actually degrade the overall throughput in proportion to the percentage of terminals that do not support
MIMO.
The LTE values show 2X2 MIMO with 1.5 bps/Hz/sector moving to 1.73 with SIC or general interference cancellation
and 4X2 MIMO. This 4X2 operation uses a simplified switched-beam approach standardized in Release 8.
Downloadable codebooks, which are being discussed in 3GPP for future releases of the standard, have potential for
future improvements beyond the 2.4 bps/Hz/sector with 4X4 MIMO, although the implementation of these types of
adaptive antenna and beamforming algorithms are based on proprietary algorithms so the gains are implementation
dependant and may evolve with field experience.
7.2. AN ANALYSIS OF ANTENNA CONFIGURATIONS FOR 4X2 AND 4X4 MIMO
An attractive base station antenna solution for LTE supporting up to four layers in the downlink is to use two
horizontally separated dual-polarized antennas. This enables a compact antenna design that can utilize both the
spatial and polarization dimensions. The amount of separation between the two antennas will have different impacts
on the potential gains of beamforming, diversity, and spatial multiplexing. Realizing these gains put conflicting
demands on the antenna separation and different choices of antenna separation will result in different system
performance profiles. The antenna size is also an important parameter from a site installation point of view. It
influences various aspects, e.g., visual footprint, wind load, and site rental cost.
19
Results from a study on the impact of antenna separation on LTE system performance are presented. By means of
system simulations, evaluations are performed to aid the understanding of the antenna separation trade-off. In
addition, empirical support to the simulation results is provided by means of comparison to results from a
20
measurement campaign.
The simulations were performed with a detailed dynamic system simulator that includes models of, e.g., adaptive
coding and modulation, UE mobility, and delays in channel quality reports. It also contains an implementation of the
21
3GPP spatial channel model (SCM) and the mutual information based link-to-system interface described in, “A
22
Fading-Insensitive Performance Metric for a Unified Link Quality Model.”
23
A simulation scenario similar to the defined 3GPP case 1 was evaluated for different configurations with dual-
polarized antennas at the BS using the closed-loop spatial multiplexing transmission mode. 3GPP case 1 refers to a
macro-cell reference system deployment type with the 3GPP SCM used for channel modeling. The network consisted
of 19 sites separated 500 m with three cells per site and an average traffic load of 4 UEs per cell. Each antenna port
21
of the BS antenna was modeled according to the BS antenna model regardless of antenna separation. The notation
Ntx_x Nrx will be used for an antenna configuration with Ntxtransmit- and Nrx receive antenna elements. For downlink
(DL), 4X4, 4X2, and 2X2 configurations, comprising one or two dual-polarized antennas at the UE and BS, are
investigated. For uplink (UL), 1X4 and 1X2 configurations comprising one vertically polarized antenna at the UE and
one or two dual-polarized antennas at the BS. (The E-UTRA standard for LTE assumes the use of at least two
24
antennas in the UE, at least as a baseline. )
Next we consider the performance impact of changing the separation of two columns of base station antennas, such
as (H) in Figure 3.
59/135
Figure 39. Downlink bit rate (left), downlink transmission rank probability (middle), and uplink bit rate
(right) as a function of the two dual bs antennas separation for the 4X4 and 1X4 antenna configuration in
the DL and UL, respectively.
The left plot Figure 39 shows normalized downlink (DL) bit rate for the 4X4 antenna configuration as a function of the
separation, given in wavelengths λ, between the two BS antennas. Three different metrics are shown: cell throughput
and the 5- and 95-percentile of the CDF of the active radio link bit rate (ARLBR), respectively. The ARLBR is the user
bit rate averaged over the time a user has been assigned resources. The bit rates have been normalized in such a
way that it is one at 1λ separation between the dual-polarized BS antennas for each percentile curve. The middle plot
shows results from the 4X4 antenna configuration of the probability of a certain transmission rank as a function of the
two dual-polarized antennas separation. The results in the left plot show that the cell throughput and cell edge bit rate
decrease as the base station’s antenna separation increases, while it is essentially constant for peak rate. There is a
benefit of a small antenna separation in this scenario since it is interference limited; hence, beamforming gains are
more important than spatial multiplexing gains. The rank statistics in the middle plot show that rank 1 and 2 are most
probable for small antenna separation. As the separation increases, the probability of rank 3 transmission increases.
Almost no rank 4 transmissions occur, since the Signal-to-Interference-and-Noise Ratio (SINR) is too low in this
scenario. Corresponding UL results for a 1X4 configuration are shown in the right plot in Figure 39. The results show
that in this case the bit rate increases (except for the cell edge bit rate at 10λ). The results show that in this case the
bit rate increases (except for the cell edge bit rate at 10λ-polarized antennas separation. The results in the left pl
60/135
Downlink Low Load Downlink High Load Uplink Low Load Uplink High Load
Figure 40. Performance summary of different antenna configurations for DL and UL for networks in high
or low load conditions.
Figure 40 shows a summary of the performance with different configurations for DL and UL in networks with high load
as well as in networks with low load. The bit rates have been normalized so that it is one for corresponding 2X2 and
1X2 results for DL and UL, respectively. Two different antenna separation are compared: 1λ and 10λ, representing
small and large separation, respectively. In the low load network scenarios shown, there are on average 0.1 UEs/cell.
The results show that for DL, a small antenna separation gives highest performance for all cases except for peak
throughput at low load. For UL, large antenna separation gives highest performance in all cases. However, most of
the UL gain in going from two to four antennas is achieved also with 1λ separation.
Figure 41. Results from full system simulation (left), single cell, single user simulation (middle) and field
trial (right) for downlink 4X4 antenna configuration.
In order to allow comparison to measurement results, Figure 41 shows throughput CDFs for a full system simulation
with an average of 4 UEs/cell, simulation of a single UE single cell (SUSC) scenario, and for the SUSC field trial
results, respectively. These results are for 4X4 configurations and each plot shows CDFs for two antenna
separations; 0.7λ and 25λ. The results are normalized to the median of the full system simulation CDF for the antenna
separation of 0.7λ. The measurements were performed using a single UE in a single cell scenario and only downlink
20
performance was addressed. In order to simulate a SUSC scenario, all intercell interference was turned off in the
simulator. In these simulations somewhat different parameter settings were used to better reflect the trial scenario,
61/135
e.g., getting a similar signal-to-noise ratio (SNR) range in simulations and trials. The purpose of the comparison is not
to reach an accurate agreement in terms of absolute performance numbers, but rather to illustrate that the relative
performance between different configurations show similar behavior.
Similar to previous results, the full system simulation shows that a small antenna separation gives the highest
throughput. In the SUSC simulation and the field trial the configuration with large antenna separation gives higher
throughput for UE positions with good channel quality. In these cases, the SNR is sufficiently high to benefit from the
additional spatial multiplexing gains offered by the uncorrelated antennas.
62/135
8. DEPLOYMENT SCENARIOS
Planning for future deployments of various smart antenna schemes requires an appreciation of the currently deployed
legacy cell sites as does discussions of general deployment issues affecting operators.
8.1. TYPICAL CELL SITE ARCHITECTURE
A typical Cell Site consists of the components diagramed in Figure 42 below. This includes the power main and any
wired backhaul facilities to the site, an AC to DC power rectifier and battery supply, and the eNodeB cabinets. These
cabinets may be located in a shelter for weather protection or they may be outdoor units mounted on cement plinths.
Backup diesel generators or other power sources may provide backup of one or more operators’ AC supplies. The
eNodeB base station cabinet may contain the RF modulators, RF Power Amplifiers, duplexing filters and lightning
arrestors and serve antennas through long coaxial feeder cables as shown on the left tower, or it may distribute the
RF components to positions next to the antennas as shown on the right figure with the use of the Remote Radio
Heads (RRHs). In this case, the eNodeB has a pair of fiber optic cables and power line connecting to the RRH that
has the radio modem, the power amplifier and filters for a sector as shown on the right tower below where one
sector’s equipment is shown.
Figure 42. Components of cell site, highlighting a single sector’s tower equipment.
A photograph of a typical site with outdoor cabinets connected directly to the antennas with coaxial cables such as to
the left is shown below in Figure 43.
63/135
Figure 43. Cell site under construction showing outdoor base station cabinets awaiting
coaxial cable connection to the tower mounted antennas.
Optionally, instead of RRHs, one may have tower-mounted Low Noise Amplifiers, antennas with remote electrical tilt
capability. On occasion, legacy base stations supporting one or another frequency band with various air interfaces will
sometimes be multiplexed with a newer base station so as share the same cable and perhaps the same antenna
system. When one antenna cannot support both frequency bands then duplex filters atop the tower may be used to
direct the various bands to their respective antennas.
8.2. CURRENT DEPLOYMENTS
One operator contributed that the percentages of various types of antenna placements are as given in Table 2. We
see here that towers and monopoles are used extensively (59 percent) while rooftops account for about 19 percent of
installations. As one might expect, rooftops are more commonly used in urban settings.
64/135
Table 2. National cell site categories (all towers include the 3 following subcategories of towers).
Cell Site Types %

Rooftop 19%
All Towers 29%
Self support 19%
Guyed Tower 7%
Utility lattice tower 3%
Monopole 30%
Water Tank 5%
Others 18%
In addition, these sites have the following numbers of antenna radomes per sector.
Table 3. Number of antennas per sector.
Antennas
(Radomes) per Percentage
Sector of Sectors
1 26.9%
2 61.0%
3 11.7%
4 0.3%
5 0.1%
The average number of these antennas (radomes) is 1.9. Additionally, the majority of the sites have 3 sectors.
8.2.1. TYPICAL TOWER TOP DEPLOYMENT
Knowing that about 30 percent of the deployments are on monopoles, it is worthwhile observing in the following figure,
the typical mechanical constraints atop a typical monopole where the so-called, “pinwheel” provides a platform for
positioning antennas, Remote Radio Heads, tower top amplifiers, LNAs, feeder cables and the like, including
workspace for installers. These pinwheels are typically placed no closer than 10 feet vertical separation. Notice the
16 feet of linear separation available for antennas sharing the same “sector face.” For an antenna configuration such
as Antenna Configurations (H) or (I) (Div-2X or TX-DIV) in Figure 44, this corresponds to as much as 11 λ at
700 MHz.
65/135
Figure 44. Tower top “pinwheel” with dimensions.
66/135
Figure 45. Common monopole during installation of antennas.
The five operators sharing the leased monopole above all use coaxial cables to connect to their antennas, and all use
similar triangular mounting platforms except for the second from the top. It uses struts such as those shone in Figure
46, below. Ropes for hoisting the antennas are still visible in this photo taken in March 2010.
8.2.2. TYPICAL MAST DEPLOYMENT
One can see in the photograph in Figure 45 above, that the second platform from the top is not a full “pinwheel” but a
cantilevered strut assembly detailed below in Figure 46. These platforms do not have the same workspace available
67/135
as the full triangular platforms shown above but they can still use Remote Radio Heads and can be equipped with
antennas spaced as far apart as those shown above.
Figure 46. Typical monopole antenna installation as seen from above and with remote radio heads
mounted on the monopole.
When mounted on a mast the cabling appears as shown below in Figure 47, where we see how the wind loading from
cables can be substantial, and how some sites have limited room for additional coaxial cables.
68/135
Figure 47. Mast with 11 operators sharing the same facility (including public safety and low power AM radio
broadcasting).
8.2.3. TYPICAL ROOF TOP DEPLOYMENT
Rooftop deployments promise much shorter cable runs as shown in Figure 48 below. However, it is often necessary
to place the cabinets indoors on a separate floor for ease of access (elevators often do not go to the roof and so a
crane would otherwise be required to place large equipment on the rooftop).
69/135
Items 1-3 of Figure 48 are three cross polarized antennas located at three corners of the roof, each placed on non-
penetrating mounts (that do not penetrate the rooftop and hence do not threaten to cause a leak in the roof.) Item 4 is
the base station cabinet and each antenna has a remote radio head mounted with it as shown in Figure 49.
Figure 48. Typical rooftop installation with three sectors. the top scale spans 5 meters and the
bottom 12 meters.
70/135
Figure 49. Antenna mount with remote radio head.
71/135
9. MISCELLANEOUS COMMERCIAL AND DEPLOYMENT ISSUES
Overarching issues of wind loading, zoning, and rental covenants apply throughout the various antenna deployment
schemes. Operators often have lease agreements dating back several years that stipulate the number of cables,
antenna dimensions, weights and loadings and so forth that would have to be renegotiated at more restrictive terms if
configurations were to be changed. While these restrictions are not universal, they can constrain the adoption of
smart antennas and 4G systems. To install a larger antenna or RRH, an operator may have to pay for a new
engineering study of the weight and wind load on all relevant structures, for example, adding considerable expense to
an upgrade.
9.1. CONSTRAINTS ON THE ANTENNA DEPLOYMENTS DUE TO COMMERCIAL CONSIDERATIONS
Standard lease agreements for the placement of antennas, cables, and wireless equipment have become increasingly
25
restrictive. For example, a typical publically available municipal lease agreement requires:
1. Setback should be 300 percent, so a 30 m tower should have a 90 m setback from the property line.
2. Panel antenna no more than two feet (2’) wide and six feet (6’) long, extending above the structure to
which they are attached by no more than ten feet (10').
3. Antenna, antenna array and support structures not on publicly owned property, which shall not extend more
than ten feet (10') above the highest point of the structure on which it is mounted. The antenna, antennas
array, and its support structure shall be mounted so as to blend with the structure to which the antenna is
attached. The antenna and its support structure shall be designed to withstand a wind force of one hundred
(100) miles per hour without the use of supporting guy wires. The antenna, antenna array, and its support
structure shall be a color that blends with the structure on which they are mounted
We have seen in other similar lease agreements restrictions specify that the overall loading on towers be restricted to
26
less than 85 percent, but leave it to operators to improve the site or mast to support any additional loads.
These restrictions on antenna form factors are further constrained by the aesthetic judgments of zoning commissions
and other municipal boards and property owners. Moreover, installation costs increase substantially if cranes are
needed when hoists are insufficient. In order to standardize configurations some tower managers have stipulated that
antennas and other equipment mounted on a mast or tower weigh in the range of 20 to 30 pounds for each such
entity, for ease of hoisting and positioning by a single climber. The range of sizes can be in the 0.75 to 1.5 foot wide
by 5 to 10 inches deep by 70 to 100 inches tall, which can be a severe limitation on the antenna scheme used. For
example, a four-column array of vertically polarized antennas spaced at λ/2, which at 750 MHz (band class 13) takes
a width about 2.6 feet.
Rooftop installations typically restrict the height of towers above the roofline, the mounting of rooftop equipment on
non-penetrating plinths (so as not to cause the roof to leak rainwater). And zoning and aesthetic considerations are
limiting the visual impact of the antennas. This impacts not only their coloration (to look like trees, for example) but
also their overall area. Even in India and China, there are field reports of having to remove eight-column antenna
arrays because of their visibility.
72/135
9.2. ELECTRICAL AND MECHANICAL TILTING OF ANTENNAS
9.2.1. EFFECTS OF MECHANICAL DOWNTILT ON SECTOR ANTENNA HORIZONTAL PATTERNS
Background
Cellular Networks achieve large capacity capabilities by re-using given frequencies repeatedly in a given system.
This concept means that the communication paths are interference limited as opposed to traditional radio systems
that were noise limited. To minimize interference, the use of sectorized antennas have been employed, each of which
provides coverage to a portion of the cell. In a three-sector arrangement, each sector antenna covers a 120-degree
pie shape that extends some distance away from the antenna site. Likewise, for a six-sector site, each antenna
covers a 60-degree pie shape.
The cellular concept presents system designers with some challenges. Ideally, each sector antenna should only
provide coverage in its 120-degree pie shaped sector so that interference with adjacent sectors is minimized. Often,
the elevation beam is tilted downward to minimize interference with other cell sites. However, in both cases, there
must be enough coverage overlap to facilitate high-speed handoffs so that calls are not dropped during the handoff
process.
Antenna designers go to great lengths to develop sector antennas that have characteristics such as fast azimuth
pattern roll off past the 3 dB points, high front-to-back ratios and excellent upper side lobe suppression on the
elevation patterns. All of these features help to minimize harmful interference.
Traditional voice networks have been shown to be fairly tolerant of interference since voice conversations are still
intelligible even with an occasional interference burst causing a few percent BER. However, with the introduction of
high-speed data networks, optimizing the desired-to-undesired signal ratios will become a prime consideration and
optimized sector antenna patterns can provide great benefits.
Network Optimization
As networks were built out, it was found that the distance covered by various sector antennas had to be adjusted
because of unwanted interference. Issues like terrain coverage, non-optimal site location and capacity driven cell
splitting are some examples. Several choices were available to accomplish this including lowering the antenna
height, changing to an antenna with lower gain or down tilting the elevation beam of the existing antenna – either
mechanically or electrically. The latter option was usually the most cost effective and has been heavily used. This
down tilting lowered the gain as viewed on the horizon since this is the area where interference takes place. As
Figure 50 shows, the angle below the horizon generally becomes very small at typical cell spacing distances. The
various tilted elevation patterns of Figure 51A and Figure 51B demonstrate how gain on the horizon – noted by the
dashed line – changes with beam down tilt. This paper defines the yellow horizontal pattern cut as the one taken
through the dashed line, while the orange azimuth pattern cut is the one taken through the peak of the elevation
beam.
73/135
Figure 50. Angle below horizon for various tower heights.
Figure 51. How gain on the horizon changes with beam down tilt.
However, there are some important differences between mechanical and electrical down tilt. Figure 52 shows the
results of a mechanically down tilted antenna where the main beam is tilted down at bore sight, tilted up at the 180-
degree point behind bore sight and not tilted at all at the ±90-degree points. Figure 52B shows the results for an
74/135
electrically tilted antenna where the tilt is the same at bore sight, the ±90-degree points and 180-degree point. This
leads to some important differences when viewing the horizontal pattern cut.
Those differences in the horizontal pattern cuts are shown in Figure 53A and Figure 53B. As expected, the
mechanically down tilted antenna’s gain is reduced at bore sight, but interestingly it is not reduced at all at the ±90-
degree point. At small values of down tilt, the pattern seems acceptable but with greater amounts of down tilt the
pattern takes on a “peanut” shaped look. This is undesirable since the purpose of down tilting is to reduce the
coverage in all directions. As the better alternative, the electrically tilted antenna with its gain is reduced in all
directions as shown in Figure 52B.
Figure 52. Differences between mechanically down tilted and electrically tilted antennas.
Figure 53. Differences in the horizontal pattern cuts.
Even though the mechanically tilted antenna’s horizontal pattern cuts look acceptable at small values of down tilt, a
subtle difference is taking place – commonly referred to as pattern blooming. In essence, the 3 dB beam width is
getting larger. Blooming patterns become important from an interference standpoint because they increase the gain
75/135
at the sector crossover points and they increase the sector overlap angle. The widely used 65-degree azimuth
antenna yields a typical 10 dB crossover gain and this has become the de facto specification for most modulation
schemes in high capacity areas.
M( )E( ) Tilt Angle Crossover
M0E0 & M0E7 ---- 17° 10 dB
M7E7 ---------------- 25° 6 dB

Figure 54B
M14E0 -------------- 29° 4 dB
Figure 54. Minimal horizontal pattern change using electrical down tilt only.
In Figure 54 the Legend Table shows the value of mechanical tilt as M=[tilt value] and the value of electrical tilt as
E=[tilt value]. Note that the normalized dark green pattern for E0_M0 and the light green pattern for M0_E7 practically
overlay each other. This substantiates the concept of minimal horizontal pattern change using electrical down tilt
only. The other normalized patterns of Figure 54 also show that when mechanical down tilt or combinations of
mechanical and electrical down tilt are employed that the blooming becomes an issue. To determine the overlap
angles, this example used a 6 dB differential as a method of comparison. This overlap can contribute to interference
in time division multiplex systems and pilot pollution in code division multiplex systems.
In the past, a rough rule of thumb emerged that allowed system designers to determine what amount of mechanical
down tilt could be considered acceptable based on blooming of 10 percent. The rule stated that an antenna should
never be mechanically down tilted more than one-half of its vertical beam width. However, this rule did not take into
account the fact that the antenna may also incorporate electrical down tilt.
Combined Mechanical and Electrical Down Tilt
Initiated by customer complaints about bad interference at a site employing both mechanical and electrical down tilt,
an investigation was opened to analyze the effects of combined mechanical and electrical down tilt. This combination
of tilts modifies what was shown in Figure 52B such that Figure 55 results. It shows that the horizontal gain is
76/135
reduced at bore sight, somewhat reduced at ±90 degrees and actually increased at 180 degrees compared to the
pattern having only electrical down tilt. Similar to mechanical-only down tilt, the horizontal pattern’s uneven gain
reduction across the sector again causes blooming to occur. The challenge was to come up with a new rule of thumb
that could predict the percentage blooming for a particular antenna given the electrical and mechanical down tilt
settings.
Figure 55. Combination of tilts.
First, range patterns were measured on different antennas for both the azimuth (through elevation bore sight) and
horizontal (on the horizon) cuts. This was done at various mechanical and electrical tilt angles. The findings showed
that the old rule of thumb was no longer applicable when electrical down tilt is employed.
Next, a mathematical model was developed that fit the measured data points and curves were generated showing the
typical amount of combined mechanical and electrical down tilt allowable for 10 percent and 20 percent blooming.
Some typical results are plotted as a function of the antenna’s elevation beam width in Figure 56.
77/135
LNX-6512 Blooming (Calc)
M0E0 0%
M4E0 3.1%
70%
M7E0 9.2%
M9E0 16.9%
60% M11E0 32.3%
M11E0 32.3%
M9E4 34.4% M0E4 0%
50% M9E0 16.9% M4E4 6.3%
M-tilt (% of VBW)
M7E4 18.8%
M7E4 18.8% M7E8 36.7% M9E4 34.4%
40% M7E0 9.2% M0E8 0%
M4E8 12.3%
30% M5E8 18.5%
M5E8 18.5%
M4E4 6.3%
M5E10 24.6% M7E8 36.7%
M4E0 3.1%
20% M4E10 15.4% M0E10 0%
M4E8 12.3% M4E10 15.4%
M3E15 39.4%
M5E10 24.6%
10% M2E15 16.7%
M0E15 0%
M0E0 0% M0E4 0% M0E8 0% M1E15 6.1%
M0 M0 M1E15 6.1%
0% E1 E1
0 0% 50 M2E15 16.7%
%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% M3E15 39.4%
E-tilt (% of VBW) 10% Blooming

20% Blooming
Figure 56. Mathematical results plotted as a function of antenna’s elevation beam width.
LNX-6515 Blooming (Calc)

70%
M6E0 40.5% M6E2 78.1%
M0E0 0%
60% M2E0 2.9%
M4E0 13.1%
50% M6E0 40.5%

M4E2 25.0%
M-tilt (% of VBW)
M4E4 46.9% M8E0 97.3%

M4E0 13.1%
M0E2 0.0%
40%
M2E2 6.3%
M4E2 25.0%
30% 10% Blooming
M2E4 10.4% M2E8 57.8% M6E2 78.1%
M2E0 2.9% M0E4 0.0%
20%
M2E2 6.3% M2E4 10.4%
M1E8 15.6%
M4E4 46.9%
10%
M0E8 0.0%
M0E0 0% M0E2 0.0% M0E4 0.0% M0E8 0.0% M1E8 15.6%
0% M2E8 57.8%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 20% Blooming
E-tilt (% of VBW)
Figure 57. Comparison mathematical results plotted as a function of antenna’s elevation beam width.
As these curves demonstrate, the green 10 percent blooming line is a function of both mechanical and electrical down
tilt. The more electrical down tilt is employed in an antenna, the less mechanical down tilt can be used to stay within
the 10 percent goal. It is interesting to see that when these two different antennas are normalized using their vertical
beam widths, the blooming percentages track very closely. This mathematical model has been used on a series of
antennas and the blooming percentages track very closely. From these plots, a new rule of thumb has been
developed for the common 65-degree azimuth beam width antennas. It states:
65° AzBW M-tilt10% Bloom = (VBW – E-tilt)/2.5
78/135
More accurate range data shows that even with no electrical down tilt the maximum mechanical tilt for 10 percent
blooming is no longer half (50 percent) the vertical beam width but that it is the vertical beam width divided by 2.5 (40
percent). When large amounts of electrical down tilt are employed, acceptable mechanical down tilt can be limited to
as little as 10 percent (or less) of the vertical beam width.
Another caution should be noted when using ultra-high gain antennas that incorporate electrical down tilt. It is
extremely important to mount them exactly plumb to minimize any possible blooming. For instance, an antenna
having 4 degrees elevation beam width and 2 degrees of electrical down tilt suffers 10 percent horizontal pattern
blooming with less than 1 degree of mechanical down tilt!
Further investigation of antennas having azimuth beam widths different than 65 degrees showed that their blooming
did not follow the rule above. After checking several different azimuth beam width models the graph of Figure 58
evolved.
Figure 58 shows that the k-factor, which is 2.5 for 65-degree azimuth models, ranges from 1.5 for 33-degree azimuth
models to 3.3 for 90-degree azimuth models. Note that these rules of thumb describe typical band-center
performance and can vary somewhat at the band edges. They also only hold true if the combined mechanical and
electrical tilts do not tilt the pattern beyond its first upper null.
Mechanical Downtilt
Factor for 10% Horizontal Blooming
Xº HBW M-tilt10% Bloom = (VBW – E-tilt)/k
3.5
3.0
k Factor
2.5
k vs HBW
2.0
1.5
1.0
30 35 40 45 50 55 60 65 70 75 80 85 90
Rated Azimuth Beamwidth (deg)
Figure 58. Variations in the K-factor which is 2.5 for 65-degree azimuth models ranges from 1.5 for 33-degree
Azimuth models to 3.3 for 90-degree azimuth models.
Recall that the new rules of thumb are most accurate near the band-center of the antenna’s specified frequency
range. They are also most accurate at the lower portion of the electrical down tilt range. In addition, the combination
of electrical and mechanical tilt should never go beyond the first upper null of the elevation pattern.
79/135
Further Findings
In addition to horizontal pattern blooming, several other important antenna specifications were seen to degrade with
mechanical down tilt. These include squint, front-to-back ratio and cross-polarization ratio. Unfortunately, these are
not easily mathematically modeled like the blooming issue. The following results from a relatively small number of
antennas have been extracted from actual measured data.
Horizontal Beam Squint
Squint is defined as the difference between the mechanical bore sight and the electrical bore sight of an antenna as
shown in Figure 59. The mechanical bore sight is defined as being perpendicular to the antenna’s back tray while
electrical bore sight is defined as the mid-point of the 3 dB beam width.
Figure 59. Horizontal Beam Squint. (Note the approximately 10-degree clockwise rotation of the beam
compared to the mechanical bore site.)
80/135
Beam Peak (Bisected at 3 dB)
Max and Min over Band vs M-tilt
(E-tilt = 0)
6
Max and Min Squint (degrees)
4 48 in X-pol 850
48 in X-pol 850
2
48 in V-pol 850
0 48 in V-pol 850
-2
-4
-6
-8
0% 20% 40% 60% 80% 100%
M-tilt (percent of VBW)
Figure 60. Beam peak versus mechanical tilt as a percentage of vertical beam width.
The previous figure illustrates how squint stays relatively constant for both vertically polarized (V-pol) and Dual-pol (X-
®
pol) 65 models when electrical tilt is varied while mechanical tilt is fixed at zero. Figure 61 analyzes the same
antenna models but now mechanical tilt is varied while electrical tilt is fixed at zero. The data is once again presented
as a percentage of the antennas’ vertical beam widths for easy comparison. It shows that squint on X-pol antennas
degrades with mechanical down tilt to the point that it can be greater than 10 percent of the antenna’s azimuth beam
width at high angles of mechanical down tilt.
81/135
Beam Peak (Bisected at 3dB)
Max and Min over Band vs E-tilt
(M-tilt = 0)
8
Max and Min Squint (degrees)
6
48 in X-pol Squint
4
48 in X-pol Squint
2 48 in V-pol Squint
0 48 in V-pol Squint
-2
-4
-6
-8
0% 20% 40% 60% 80% 100%
E-tilt (percent of VBW)
Figure 61. Varied mechanical tilt with electrical tilt is fixed at zero.
Sector Power Ratio
Sector Power Ratio (SPR) is another measure of an antenna’s ability to minimize unwanted interference in a cellular
network. Figure 62 graphically shows the concept along with the equation used to calculate SPR. In essence, it
compares the undesired RF power outside of the sector to the desired RF power within the sector and expresses it as
a percentage. Excellent antenna designs provide SPRs as low as 3 percent to 4 percent, while traditional dipole or
patch element designs can yield SPRs approaching 8 percent.
Figure 63 shows that similar to squint, SPR remains relatively constant for both V-pol and X-pol models as the
electrical tilt is varied while keeping the mechanical tilt at zero. In Figure 64, SPR is shown to degrade significantly for
all types of antennas when mechanical tilt is varied while keeping electrical tilt fixed at zero.
82/135
300 120°
∑ PUndesired
SPR(%) = 60
60
× 100
∑P
300
Desired
Figure 62. Graphical representation of Sector Power Ratio (SPR).
Sector Power Ratio vs E-tilt

M-tilt = 0
16
14
SEctor Power Ratio (%)
12
10
48 in X-pol 850
8 96 in X-pol 850
48.5 in V-pol 850
6
0
0% 20% 40% 60% 80% 100%
Figure 63. SPR remains relatively constant for both v-pol and x-pol models as the electrical tilt is varied.
83/135
Sector Power Ratio vs M-tilt
E-tilt = 0
16
14
Sector Power Ratio (%)
12
10
48 in X-pol 850
8 96 in X-pol 850
48.5 in V-pol 850
6
0
0% 20% 40% 60% 80% 100%
Figure 64. SRP degrades significantly for all types of antennas when mechanical tilt is varied which keeps
electrical tilt fixed at zero.
Front-to-Back Ratio and Cross-pol Ratio
When either (or both) of these two specifications degrade, system performance can be affected – but in different
ways.
Front-to-back ratio (F/B) is traditionally measured by comparing gain at bore sight to gain at point 180-degree behind
bore sight as shown in Figure 65A. Some system specialists have taken this concept to another level by
characterizing the F/B ratio over some angle around the 180-degree point – such as 180 ±30 degrees. Either way it is
a measure of unwanted interference behind the desired sector.
Cross-pol Ratio (CPR) as shown in Figure 65B is a measure of the de-correlation of the two polarizations used in a X-
pol antenna – one at +45 degrees and the other at -45 degrees. Within the sector, having gain de-correlation of at
least 10 dB assures good uplink diversity in a multipath environment. Most cellular systems depend on some sort of
receive diversity at the cell site to balance the uplink path from the lower power cell phone devices with the downlink
path from the higher power cell site transmitters. Lack of good uplink diversity in essence shrinks the coverage
distance of the site.
It should also be noted that for X-pol models, the total F/B power is the sum of the co-pol energy plus the cross-pol
energy.
84/135
--Co-Polarization
--Cross-Polarization
(Source @ 90°)
-5
-10
-15
-20
-25
-30
-35
-40
F/B Ratio @ 180
0 dB – 25 dB = 25
Figure 65A. Front-to-back ratio measured by Figure 65B. Cross-pol ratio is a measure of the
comparing gain at bore site to gain at point 180 de-correlation of the two polarizations used in
degrees behind bore site. an x-pol antenna — one at +45 degrees and the
other at -45 degrees.
Figure 65. Front-to-back and Cross-pol ratio.
To demonstrate the huge amount of degradation that can happen to both F/B ratio and CPR when mechanical down
tilt is applied to an antenna already having a large amount of electrical down tilt, the measured patterns of Figure 66
are presented. The patterns are for a typical 4 foot, 65-degree antenna employing 15-degree electrical down tilt along
with a 5-degree mechanical down tilt. The co-pol pattern is shown in blue and the cross-pol pattern is shown in red.
With a vertical beam width of ~16 degrees at 850 MHz, the tilt combination is well beyond even the 20 percent
blooming curve of Figure 56. In fact, the horizontal beam width is in the neighborhood of 160 degrees or
approximately 250 percent blooming!
The F/B ratio has degraded to approximately 18 dB at 180 degrees and the cross-pol is actually worse than the co-pol
pattern. Taken over a ±30-degree angle around 180 degrees this F/B method measures only 8 dB!
Finally, the CPR over the desired sector degrades to only 5 dB at the sector edge – far short of the 10 dB expectation.
85/135
Figure 66. Demonstration of the huge amount of degradation that can happen to both f/b ratio and cpr when
mechanical down tilt is applied to an antenna having a large amount of electrical down tilt.
Conclusions Concerning Electrical and Mechanical Downtilt
The industry has always realized the coverage and interference compromises associated with mechanically down
tilting sector antennas and a rule of thumb commonly used to minimize these issues; however, it did not incorporate
the use of electrical down tilt.
This sector analysis goes into more detail by showing how various parameters including horizontal beam width,
squint, Sector Power Ratio, front-to-back ratio and cross-pol ratio are all affected by mechanical down tilt.
It presents new rules of thumb that limit blooming to approximately 10 percent when combined mechanical and
electrical down tilt are employed. Also included is measured data showing how the other important parameters
®
mentioned above – especially for the popular Dual-pol (X-pol) models – degrade when mechanical down tilt is used.
As the industry evolves toward data systems supporting higher and higher data rates, more attention must be paid to
these parameters, which can severely degrade overall network performance, and the importance of using only
electrical down tilt.
86/135
9.2.2. EFFECTS OF INCORRECT ANTENNA INSTALLATION
Background
As was stated in the previous section, cellular networks re-use frequencies repeatedly again to achieve the capacities
demanded by today’s customers. In order to achieve optimum performance the sectorized antennas used at each cell
site must be installed according to the design engineer’s specifications. If they are not installed correctly, network
27,28
performance will suffer.
Real World Issues
Many installers who have the dangerous job of climbing cellular towers have not been properly trained on the
importance of correct antenna installation. Some common issues are noted below:
The antenna support pipe must be plumb. Often installers will mount an antenna such that it is parallel to the support
pipe. Of course, if the support pipe is not plumb, the antenna will not be plumb. This can produce mechanical down-
tilt, mechanical up-tilt or side-to-side skew depending on the support pipe’s orientation. The correct procedure is to
use a digital inclinometer to plumb the antenna in both planes using the adjustable mounting brackets supplied with
most sector antennas.
The azimuth pointing direction is not correct. The design engineer specifies a geographic pointing direction or
heading for each antenna to insure optimum network performance. However, if care is not exercised, a couple of
issues can cause serious problems. First, magnetic compasses do not read correctly when used near a large amount
of metal and since towers use large amounts of metal, readings taken at or near the antenna will be faulty. Usually it
requires someone on the ground – away from metallic structures – to establish the correct pointing direction and
coordination with the installer on the tower to then point the antenna correctly. The second issue that untrained
installers encounter is the magnetic declination angle between magnetic and geographic North. In some areas,
differences greater than 15 degrees exist. To complicate things even further, the declination changes from year to
year so it is important that installers know where to access the latest information.
Azimuth adjustment tools are available on the market that makes use of relative bearing references to known large
objects, such as churches, chimneys and silos, rather than magnetic compass bearings. Relative bearings can be
obtained from a map and then used as a reference for the azimuth adjustment tool while adjusting the antenna
azimuth direction.
87/135
Figure 67. Azimuth Adjustment Tool.
The mounting hardware must be assembled correctly. Antennas either are supplied with mounting kits or have them
available as a separate purchased part. If the installer does not follow the installation instructions, some of these
mounts have been known to change position under high wind conditions. The network will suffer until another tower
climb is initiated to repair such a faulty installation.
9.3. PASSIVE INTERMODULATION (PIM) SITE CONSIDERATIONS
Background
Intermodulation is defined as the mathematical mixing of two or more desired RF signals to generate a family of
29
undesired RF signals. For years, this phenomenon has been well understood in active components such as power
amplifiers and preamps. With the introduction of duplexed, high data rate systems the industry has become focused
on this same intermodulation happening in passive RF components – especially those that share both downlink (high
power transmit) and uplink (very low power receive) signals. If undesired passive intermodulation (PIM) products are
produced in the receive frequency band of an FDD system, there is no way to eliminate them by filtering signals and
they will interfere with or degrade the desired signals coming from user equipment (UE).
Discussion
Intermodulation can take place whenever multiple RF signals encounter some type of non-linearity. It can take place
in the RF path or beyond the RF path after the signals are radiated from the antenna. Things such as rusted tower
joints or guy wires as well as rusted or corroded equipment on a building top are all examples of PIM generators.
Loose nuts and bolts or rivets are other examples.
All RF path components such as connectorized transmission lines and jumper cables, antennas, diplexers, tower
mounted amplifiers (TMAs) etc., have some inherent degree of non-linearity. The industry’s challenge is to develop
designs that suppress the unwanted intermodulation to acceptable levels. The accepted standard for antenna PIM
uses two 20 watt transmitters (A and B) and states that any third order intermodulation products (2A-B or 2B-A) must
be at least 150 dB below the two transmitter carriers (-150 dBc).
Another type of third order intermodulation product takes the form of A + B – C where three desired signals are now
involved. It also opens up the possibility that different frequency bands and blocks can be involved. This type of PIM
can happen when diplexers are used to minimize the number of transmission lines at a site. For instance if a US
88/135
operator had PCS blocks B and F and decided to diplex them with the Upper 700 C block, there is a possible PIM hit
30
(756+1973-1952=777). A calculator is available that allows designers to check for PIM problems when various
band(s) and block(s) are combined. It is strongly recommended that these combinations be avoided since they are
very difficult to eliminate in practice.
Other less complex components like connectorized transmission lines and jumpers normally are specified to have
better PIM suppression in the range of -160 to -163 dBc. This is important since in the complete RF path there will be
a number of PIM generators and the system PIM will depend on all of the contributors. Each PIM contributor is a
vector having amplitude and phase so they can combine in various ways. The typical result is a random combination
of the individual vectors, but under worst-case conditions, they could all combine in-phase.
31
Testing Considerations
There are several test equipment companies that have developed rugged, high performance, portable PIM test
equipment. Most can test either throughput PIM or reflected PIM. Since the industry is concerned about PIM coming
back into the receivers, most components are tested for reflected PIM. This equipment must be complimented with a
good low-PIM 50-ohm test load and a good low-PIM jumper cable before meaningful testing can be done. Testing
two-port devices like coax cables for reflected PIM is relatively straightforward and simply requires that the low-PIM
load be connected to the output port. Be sure to correctly torque all RF connectors. However, testing antennas for
PIM is a much more meticulous process. For best results, the antenna should be tested outdoors facing skyward with
no metal within 15 to 20 feet. Testing indoors, even in large warehouses, should be avoided as these environments
have shown to produce incorrect results. In addition, no personnel should be close to the antenna during testing.
Items like keys in someone’s pocket or a cell phone on their belt can cause undesired PIM. Again, applying the
correct torque to RF connectors is an important step.
For field-installed connectors on transmission line or jumper cables, use of the recommended cable preparation tools
is very important. Follow the manufacturer’s instructions closely to insure the best results. Smooth flared surfaces
with no burrs on either the center or outer conductor is required for low-PIM connectorization.
Finally, use only 7-16 DIN style connectors on all RF path components where possible. It is important that the
connectors being used are approved for the cable brand being used to ensure the correct fit and sealing. Also, insure
that the installers have the correct 25 Nm (222 inch-pound) torque wrenches.
9.4. INDEPENDENT ANTENNA TILT OPTIMIZATION BY AIR INTERFACE
It is common for operators that have multiple air-interface technologies in their network to elect to optimize the
services independently. For example, an operator that is utilizing GSM in the PCS band and UMTS in the AWS band,
may elect to use two separate variable electrical tilting antennas rather then diplexing the services on one antenna.
This approach is chosen even though the use of one antenna could have advantages for wind loading, leasing costs,
and feeder cable reduction. Often quad antennas are used for independent tilt to minimize the radome count on the
tower. Likewise an operator who deploys a multi-band antennas using GSM on the cellular band and UMTS on the
PCS band can implement independent tilt to optimize the two services and associated frequency band separately.
This independent tilt is particular important for muti-band antennas given the relative differences between 1 GHz and
2 GHz for RF propagation, antenna gain and vertical beam width.
As described in Section 6, novel new active antenna arrays are emerging that allow for the phase and amplitude
weighting of signals on all antenna elements, not just columns of elements. This enables beam forming and MIMO
operations in both the azimuth and now the elevation direction. Moreover, this can be done on a per carrier bases, or
even on a per user basis.
89/135
With proper support in the baseband processing, these Active Antenna Arrays can conceivably tilt different air
interfaces on different carriers in the same band by different amounts. This can be particularly attractive for
introducing LTE for hot spot coverage, where GSM/GPRS/EDGE or WCDMA/HSPA is tilted up for broad coverage
while LTE is focused down at the local vicinity.
9.5. REMOTE RADIO HEADS FOR MIMO
Remote radio heads (RRH) also known as remote radio units (RRU), are currently being deployed not just for new
technologies (such as LTE) but also in new and replacement infrastructure using older technologies (2G, 2.5G, 3G),
primarily for reasons of lower capital and operating expenditures. The deployment of remote radio heads is
accelerating as system operators become more comfortable with the reliability of the equipment, and are prepared to
move the transmit and receive electronics close to the antenna(s) in each sector. In some cases, where there is not
high confidence in the reliability or for other reasons, RRHs are deployed at ground level and traditional coaxial
feeders are used to connect to the antenna(s). These ground level remote radio heads are also called transmitter-
receiver duplexer units or TRDU, also known as a Radio Unit (RU).
9.5.1. REMOTE RADIO HEADS AND TRANSMITTER-RECEIVER DUPLEXER UNITS
The major components of an enhanced Node B (eNB) include the transmitters and receivers (transceivers), power
amplifiers, duplexers, antennas, low noise amplifiers (LNAs), operations and management (OA&M) controllers,
antenna controllers, and baseband processing units. Traditionally the majority of the components of an eNB reside in
a shelf at the base of the tower. RF cables are used to connect to the antennas at the top of the tower.
Transmitter-Receiver Duplexer Unit
The TRDU is a shelf-mounted module that integrates several functions of the eNB and is typically located near the
base band unit in the eNodeB cabinet. In a typical implementation, the TRDU includes the transceivers, power
amplifiers, antenna controllers, and duplexers. The TRDU is connected to the baseband processing unit via a high-
speed optical serial interface and connected to the antenna via RF cabling. Forced air-cooling is employed in the eNB
shelf
Remote Radio Head (RRH)
A RRH, also known as a Remote Radio Unit (RRU), incorporates much of the same functionality as a TRDU.
However, the RRH is a pole-mountable, weatherized, self-contained module that uses natural convection cooling. In
a typical installation, the RRH is installed close to the antennas at the top of the tower with minimal RF cabling.
Figure 68 shows the block diagram of a TRDU/RRH with two transmitters and four receivers. The interface to the
baseband processing unit is a high-speed optical interface (here a CPRI interface is shown). The interface board in
the Figure incorporates some of the control functions such as antenna control (AISG) and alarming.
90/135
Figure 68. Block diagram of a RRH/TRDU.
Figure 69. Examples of Remote Radio Heads with two transmit and receivers and with one transmitter and
two receivers.
91/135
9.5.2. REMOTE RADIO HEAD ADVANTAGES AND CONSIDERATIONS
Remote radio heads have several advantages over traditional Node-B architectures.
• Site Footprint--Since they are mounted directly on a pole, they require little or no site footprint. This can
significantly reduce capital expenditure by reducing site planning and maintenance costs
• Power Consumption--The proximity of the transmitter to the antenna can minimize feeder cable losses and
thus improve the overall base-station power consumption.
• Receiver Performance--The proximity of the receivers to the antennas also improves the sensitivity and
noise figure of the receivers.
• Flexible Network Coverage--Limited site space availability in some areas can prevent the installation of a
traditional BTS. An RRH can fill such coverage holes.
• MIMO Capability--The gains, phases, and delays of each individual carrier and/or the composite signal in
each transmitter and receiver in a RRH can be digitally adjusted. Each RRH also has a measurement
receiver that is used to calibrate the digital pre-distortion algorithms. The measurement receiver can be used
to make accurate measurements of the gains, phases, and delays of each transmit path. These features
enable easy implementation of MIMO and SIMO processing architectures with RRHs.
• Daisy-chaining RRHs--Multiple RRHs can be daisy-chained using the optical interface and programmed to
operate in a coordinated manner. This feature can be used to increase site capacity and coverage. It can
also enable higher order MIMO and SIMO systems. For example, an RRH with two transmitters and two
receivers can support a 2X2 MIMO system. Daisy chaining two such RRHs can enable a 4X4 MIMO system.
• Reliability--RRHs and TRDUs are inherently more reliable than traditional eNBs because of the integration of
several subsystems in a single functional unit. RRHs, which are convection cooled, are less reliable than
TRDUs, which are cooled using forced air. Innovative thermal designs and highly efficient amplifier designs
have significantly improved the reliability of RRHs.
• Installation--A typical RRH can weigh 15-20 kg. Rooftop installations of RRHs are simple and
straightforward. Tower-tops installations are more complex and require planning for tower climbs.
• Weatherization--Unlike a sheltered shelf RRHs are exposed to the elements. Damage, performance
degradation, and failures can result from improper mechanical design. Therefore, careful thermal and
mechanical design is required to ensure proper weatherization.
• Battery Back-ups--Though an RRH can be installed at the top of a tower, battery back-ups will still be
required for emergency operation. Most battery back-up systems will require some site footprint. Innovative
solutions for the installation of battery cabinets near the RRHs are now available.
9.5.3. REMOTE RADIO CONFIGURATIONS
Currently, RRHs are being deployed in multiple frequency bands, across several power levels and in various
configurations.
92/135
Frequency Bands
UMTS (Node B) RRHs are commonly deployed in 2100 MHz and 900 MHz. RRHs to support the LTE trials in North
America have been deployed in the 700 MHz and 800 MHz bands. Several LTE deployments in 2600 MHz are also
underway. Additional frequency bands with early LTE deployments include 1500 MHz in Japan and 2300 MHz and
1.8 and 2.3 GHz in China with TDD-LTE installations.
The bandwidth of operation of the Remote Radio Heads is typically limited to 20 to 30 MHz in which multiple carriers
can be located.
Power Levels
Macro cell installations with transmitter powers ranging from 10 Watt – 60 Watt are most common in initial
deployments. Micro cell installations with power levels per transmitter ranging from 1-5W shall follow the macro cell
installations. Multiple transmit ports typically divide the overall transmit power so there are 1 TX x 60 Watt and 2 TX x
40 Watt units as well as 4 TX x 10 Watt units.
Transmitter-Receiver Branches
At 700 MHz a 2X2 configuration (two transmitters and two receivers) is most common. A 2X4 configuration is
preferred at higher frequencies like 2600 MHz, 2300 MHz, 1500 MHz, 1800 MHz, 1900 MHz, and 2100 MHz. As
described earlier, multiple RRHs can be combined to form other configurations such as 4X4 and 4X8. At 2300 MHz,
an 8X8 installation for TDD-LTE is expected to be widely deployed.
Air Standards/Multi-Standard
Currently RRHs are widely deployed in all major air standards such as GSM, CDMA2000, UMTS, WiMAX, and LTE.
Most RRHs will be capable of supporting multiple air standards in one unit. Common multi standard implementations
include LTE+UMTS, LTE+CDMA2000 or LTE+UMTS+GSM. An LTE+CDMA2000+UMTS configuration is expected to
be less common. RRHs shall be software configurable and capable of supporting any carrier combination with these
multiple standards.
Interface Standard
The optical high-speed serial interface between the baseband processing unit and the RRH/TRDU typically use the
Common Public Radio Interface (CPRI) standard or the Open Base Station Architecture Initiative (OBSAI) standard.
Current RRHs support data rates up to 6.144 Gbps. It is expected that the next generation of RRHs will require line
rates up to 12.2288 Gbps or higher.
This data rate on the high-speed serial interface reflects the maximum bandwidth of the spectrum transmitted and
received by the Remote Radio Head/Unit as well as the number of branches, and the number of units that may be
daisy chained with one optical high speed interface.
FDD vs. TDD
Most RRHs deployed support FDD (Frequency Division Duplexing). However, RRHs also support TDD (Time
Division Duplexing) for WiMAX (licensed and unlicensed bands) and LTE (2300 MHz).
Efficiency
The efficiency of an RRH is defined as the RF output power delivered divided by the DC input power. At macro power
levels, the power amplifier is the largest contributor to the overall power consumption of the RRH. This is not to say
93/135
that the power consumption in the other subsystems is negligible. The power consumption in the digital subsystem
can be significant (20-30 W) and this is especially critical for micro RRHs.
9.6. FEEDER CABLE FOR REMOTE RADIO HEADS
A perennial challenge in placing antennas at appropriate heights has to do with the costly, bulky and lousy coaxial
cables. There are often many tens of meters of cable connecting the antennas and the radio transmitters, which are
often located in climate-controlled enclosures on the ground near a tower base or in a central location either on the
roof of a building or on a lower floor.
So-called “hard line” coaxial feeder cable is typically 7/8” to 1-5/8” or even thicker is used for long straight runs, such
as up a tower, with more flexible and thinner jumper cables making shorter connections to the antenna itself on one
end and to a lightning arrester and the base station on the other end. These cables along with their connectors,
lighting arresters and sometimes combiners, often contribute 2.5 to 3.5 dB of RF power losses, requiring an amplifier
twice as powerful (and consuming twice as DC power) as strictly required at the antenna. In addition, these cables
add weight and wind load to a tower or mast installation.
In contrast is a remote radio head installation. The remote radio heads use fiber optic cable, either single-mode or
multimode, for the transmit and receive signals, and they require power of course. The power is typically 48V DC, but
some equipment is AC powered. Installed systems to date have used separate fiber and power cables, often installed
in conduit for mechanical protection and electro-magnetic shielding against lightning strikes.
Here we compare the relative weight and wind load of Remote Radio Heads (RRHs) as opposed to running coaxial
cables up a mast. For example, we consider the weight and wind load for a 4 branch per sector installation on a 30 m
2
mast . Using the common 1-1/4” LCF114 cable, three sectors would contribute over 310 kg of mass on the tower or
680 pounds. Cable strain relief mounting hardware placed the recommended 4 feet and cable hoist add a minimum of
20 kg depending upon the type of hardware used. The sum total static load on the 30 m tower is therefore at least 330
kg (727 lbs). A remote radio head, on the other hand, typically weights about 17 kg but with mounting hardware,
power and fiber takes about 25 kg, which is 13 times smaller than the coaxial cables that the RRH can replace. Two
branches per sector would of course have half this weight or 155 kg and would mean that the 2 cables per RRH would
weigh over 6 times that of the RRH.
Cables are typically run up a mast along a cable run with stress relief clamps that hold the cables adjacent and
parallel to each other. To reduce wind load, the cables are run up a tower in a 2 x 6 cable arrangement with space
between the cables. Consequently, the maximum wind load is when the 2 rows of 6 cables are facing the wind with an
2
effective area of 0.234 square meters per meter of height, 7 m for the full 30 m height.
The IEC 721-3-4 standard indicates a peak wind burst of 50 m/sec (180 km/hour) be considered; corresponding to the
3-second gust wind speed measured at 33 ft. above ground. [IEC72134]
The EIA/TIA-222-G standard [TIA222G] prescribes tables and equations for calculating the wind load as a function of
height above the ground, so that the relative torque and twist on a mast can be evaluated. Assuming the default
values of Topographic and Exposure categories corresponding to a flat field, the force on the cable versus 3 RRHs
are shown in Figure 70. The wind velocity varies with the height and so the force is shown for every 1-meter section of
cable.
2
Civil and structural engineers make a distinction between a self-supporting tower such as monopole while a mast uses guys or stays anchored
into the ground for lateral support. Masts require an extended area surrounding them to accommodate the stay blocks. Towers are more commonly
used in cities where land is in short supply.
94/135
Figure 70. Force on the cable vs. remote radio heads on a typical mast. this is assuming the 1-
1/4” cables are stacked two by six. the remote radio heads ranged from 12 kg to 20 kg and from
2 2
0.12 m to 0.18 m .
The weight and moment (torque) resulting from this wind load (described above) integrated over the 30 meters of
mast height are tabulated below:
Twelve 1-1/4” 6 Typical 3 Typical Remote

LCFS114 Coax Cables 2xpol Antennas Radio Heads
For 4 branches/sector
Weight 330 kg 2x3x 15 kg = 90 kg 3x 17 kg = 51 kg
Moment 136 kNm 120 to 170 kNm 15 to 22 kNm
2
The representative antenna had an area of 0.48 m and the Remote Radio Head (RRH) was exemplified with an area
2
of 0.118 to 0.18 m . The moment is generally proportional to the area and to the square of the wind velocity. Keep in
mind, too, that 2 branches per sector would amount to half the weight and as half the moment.
In addition, ice can accumulate on the cables and hardware, doubling the weight and adding to the effective area. In
Table C.1 of the TIA-222-G standard, the impact of two different thicknesses of ice is prescribed for a 9-antenna
configuration along with nine 1-5/8” cables configuration. This is a composite of all the antennas and cables together.
(Note that the cables here are larger than the earlier example.)
95/135
Table 4. Example wind load for a nine-antenna (and nine-cable) tower with ice loading.
ICE ICE
No Ice
T ≤ 0.5” 0.5” ≤ T ≤ 1.5”
Eff. Area Weight Eff. Area Weight Eff. Area Weight

2 2 2
9 antennas 6.5 m 337 kg 7.9 m 337 kg 10.2 m 693 kg
9 1-5/8”
2 2 2
Scaled 6/9 4.3 m 225 kg 5.3 m 225 kg 6.8 m 462 kg
Moment 182 kNm 222 kNm 285 kNm
The important comparison is that remote radio heads have less than 1/13 the weight and 1/6 to 1/9 the moment in
high wind bursts. Icing and larger cables such as 1-5/8 inches make the comparison even more strikingly in favor of
RRHs.
Both self supporting and guyed masts are susceptible to ice and wind, but of course monopoles with cables inside the
single hollow tube, shield the cables from icing and wind, but in monopoles, access to the cables is more difficult and
constrained by size limits as all must fit through access holes such as those in the following photo.
Figure 71. Photograph of the constrained access panel for running coaxial cable into a monopole.
96/135
In those cases where the lease agreements on the mast charge explicitly by static weight or wind load, the RRHs
have a clear advantage, but the physical size of the RRH (though smaller than typical antennas) is occasionally an
issue. On the other hand, maintenance and upgrades to an RRH can also present some difficulties due to the
required tower climb. Therefore, making the RRHs robust and “future proof” with versatile software is a priority.
Moreover, when placing an RRH next to the antenna, the 2.5 to 3.5 dB of cable loss is reduced to less than 0.5 dB,
which can be used to either increase the transmitted power by 3 dB (doubling the power) or it can be used to reduce
the power requirements on the amplifier, thus reducing the power consumption of the base station. When multiple
carriers can be configured within the operating bandwidth of the RRH, the reduced RF losses can potentially permit
about twice as many carriers at their full power.
While Remote Radio Heads (RRH) can be placed either near the antennas or on the ground with traditional cables
running the distance to the antennas, the most appropriate configuration is unclear. Active equipment such as RRHs
and Tower Mounted Antennas require “tower climbs” to repair or replace and requires that other sectors be turned
down when field service engineers are working nearby. Upgrades to add additional power amplification or bandwidth
beyond that originally provided by the RRH may require tower climbs to upgrade services, and some lease
agreements require renegotiation when equipment of a different size/weight/appearance is installed.
9.6.1. HYBRID FIBER/COPPER REMOTE FIBER FEEDERS
Now available are armored cables, remote fiber feeders (RFF), combining the fiber and power elements in a single
construction. A typical cable is shown in Figure 72. The cable consists of two fiber cable subunits each containing
either 2, 4 or 6 tightly bundled fibers, which can be either single-mode or multimode. These are stranded around a
central strength member along with three power conductors (positive, negative, ground). Around this stranded core is
a water-blocking tape, ripcord, outer shield/armor and the outer jacket. The conductor sizes are either 8 AWG or 10
AWG, depending on the power requirements of the Remote Radio Head Units, and the outer jacket can be either
outdoor grade black polyethylene, for tower and monopole installations, or a low-smoke zero halogen compound for
building installations.
97/135
Figure 72. Hybrid Fiber/Copper Cable.
These cables are typically installed one per sector between the main unit and the remote radio heads. Other
configurations are possible, however, depending on the power consumption of the remote units and the associated
voltage drop across the remote feeder. For example, a single cable might feed a remote unit at the top of a tower,
which contains the radios for all three sectors. These would then connect to the antennas with regular coaxial cables.
Multiple fibers allow for flexibility, redundancy and future-proofing of installations. Any or all of the fibers in one or
both fiber subunits can be terminated with appropriate fiber connectors. Alternatively, the appropriate fiber and
copper “break-outs” can be made and some fibers left to be terminated at some later date.
These cables, being a stranded construction, also allow the possibility of including additional elements such as an
RET control cable, or an alarm cable, in custom designs.
9.6.2. ASSEMBLIES
Because the sizes are so small, correct termination of fiber optic cables requires more specialized tools and
equipment and a higher skill level than does termination of coaxial cables. For this reason, fiber cables used with
remote radio heads up to now have usually been pre-terminated assemblies. It is expected that the same will be true
for the remote fiber feeders, at least for initial deployments. However, as this technology becomes more widespread
there will likely be a requirement for field termination of fiber cables.
A typical breakout and termination scheme is shown in Figure 73. The outer shield and jacket are stripped back to the
required lengths (say 0.5 m at the RRH end and 1 m at the main unit end) to expose the power conductors and one or
both fiber subunits. The power conductors can be enclosed in heat-shrink or other tube material, and the main
breakouts are sealed with epoxy and heat-shrink, which provides mechanical support and water blocking. The tightly-
98/135
bundled fibers are individually broken out from the subunit cable a few inches from the ends, and terminated with
suitable connectors per the equipment interfaces. Dual LC connectors seem to be the most commonly used currently.
Figure 73. Breakout and Termination Scheme.
9.6.3. LIGHTNING PROTECTION
The outer shield on the Remote Fiber Feeder (RFF) cable provides basic support and mechanical protection, but also
acts as a lightning shield. Power cables for outdoor installations often require electro-magnetic shielding outside the
conductors to reduce the induced transients on the conductors and minimize the possibility of damage to the
electronic equipment, even though this may include integral or added surge suppression. The shields on the RFF
feeders will withstand simulated strike currents in excess of 100 kA with no damage to the cables. This is also a
typical value for coaxial cables.
9.6.4. ACCESSORIES
Hoisting grips, hangers and ground straps are available for the remote fiber feeder cables, and continue to be
developed. Typical accessories are shown in Figure 74. These are similar to, and in some cases identical to,
standard accessories for coaxial cables with which the installation community is familiar.
99/135
Figure 74. Accessories.
9.6.5. ADVANTAGES OF REMOTE FIBER FEEDERS
There are a number of advantages in using the combined fiber/copper remote fiber feeders, compared with the
separate fiber and copper cables used up to now.
• Lower total material and installation costs. No conduit required.
• Ease of handling. One cable per sector. Ease of routing.
• Environmentally and mechanically protected.
• Reduced tower loading.
• Reduced tower real estate cost.
• Standard coaxial accessories.
100/135
9.7. CO-SITING OF MULTIPLE BASE STATIONS AND TECHNOLOGIES
9.7.1. SEPARATE FEEDERS FOR EACH SYSTEM
In an ideal world, each base station and air interface technology deployed on a cell site would have its own feeders
and antennas. This would have several benefits:
• Individual optimization of antenna pattern, azimuth direction, and down tilt for each service.
• Minimum RF path loss and mismatch.
• Reduced concerns for interference and intermodulation between systems.
• Maintenance work on one system need not impact other services.
However, a number of constraints influence the site architecture and reduce the system designer’s options. In most
situations, factors that must be considered include:
• Zoning restrictions on the quantity and design of antennas and other equipment.
• Limits on weight and wind loading of feeders and tower mounted equipment.
• Cost savings opportunities on capital expenditures, lease costs, and installation labor.
• Time to market requirements for rollout of new services.
A variety of co-siting solutions have been developed to address these constraints while enabling efficient operation of
multiple technologies in a shared architecture. Today, these solutions are widely used within major carriers’ networks
and competing operators have also joined together to reap the benefits of sharing site equipment.
101/135
X X X X X X X
X X X
X X X X X X X
Antenna
X X X
X X X X X X X
700 850 1900 850 1900 700/850 1.7-2.2
Jumper
850 1900 850 1900 LO HI LO HI

DIPLEXER DIPLEXER DIPLEXER DIPLEXER
COM COM COM COM
Jumper
Feeder
Hatch
Jumper
COM COM COM COM

DIPLEXER DIPLEXER TRIPLEXER TRIPLEXER
850 1900 850 1900 700 850 1.7-2.2 700 850 1.7-2.2
Jumper
T/R1 T/R2 T/R1 T/R2 T/R1 T/R2 T/R1 T/R2 T/R1 T/R2 T/R1 T/R2 T/R1 T/R2 T/R1 T/R2
Base
LTE-700 GSM-850 UMTS-1900 Station GSM-850 UMTS-1900 LTE-700 GSM-850 UMTS-1900
Figure 75. Multiband sector Figure 76. Shared feeders Figure 77. Shared feeders and
with separate feeders. using diplexers (crossband antennas using triplexers and
couplers). broadband antennas, diplexers.
9.7.2. SHARED FEEDERS—FREQUENCY MULTIPLEXING
Feeder cables are generally suitable for any frequency band and air interface technology and can therefore easily be
shared. By combining services at the foot of the tower and again separating them just below the antennas, the benefit
of individual antenna optimization is retained while reducing the number of feeders. Existing feeders can be used for
additional frequency bands, helping accelerate rollout of new services.
Two or more frequency bands are combined or separated using Crossband Couplers. According to the number of
paths combined, they are also known as Diplexers, Triplexers, etc. When the frequency separation between the
bands combined is relatively wide, such as between 700 MHz to 1000 MHz, 1700 MHz to 2200 MHz, and/or 2400
MHz to 2700 MHz, crossband couplers can be compact, low cost devices while introducing almost negligible loss and
mismatch. When combining more closely spaced bands, for example a 700 MHz band with another in the 850 MHz
range, the components need to become somewhat larger and more complex in order to maintain acceptable
performance.
102/135
Figure 78. Diplexer.
Figure 79. Triplexer.
9.7.3. SHARED ANTENNAS
Where individual antennas for each band and service cannot be installed, multiport and multiband antennas are used.
This typically restricts the azimuth angle to be the same for all bands while beam tilt can be set independently for each
antenna port or pair of ports. Crossband couplers or dual band TMAs are most often used to separate the bands
below the antenna. However, multiband antennas (mainly dual band) with built in crossband couplers are also
available on the market. These provide a very clean installation but with the disadvantage of not being able to support
a TMA on one of the frequency bands.
103/135
Using broadband antennas specified for, e.g., 1700 MHz to 2200 MHz, more than one band and service can be
combined on a feeder and connected to a single antenna port. Antenna selection and adjustment may then involve a
compromise that is acceptable albeit less than optimal for each service sharing the antenna. Technologies are
available that allow individual beam steering but they come at additional cost and complexity.
9.7.4. SHARED FEEDERS AND ANTENNAS—SAME BAND COMBINING
When two or more services or operators share the same license band, regular crossband couplers cannot be used. In
this situation, other solutions are available as described below. Same band combining is also often used with a single
service to combine base station ports or cabinets when the channel count exceeds the capacity of a single unit. In
Frequency Division Duplex (FDD) wireless systems, same band combining always involves distributing the uplink
(RX) signals from shared antennas to multiple receivers and may also include combination of multiple base station
transmitters to an antenna for the downlink (TX).
X X X X
X X X X
X X X X
X X X X
X X X X
1900 1900 1900 1900
CABLE CABLE
LOAD LOAD
ANT ANT
LLC-DUPLEXER LLC-DUPLEXER
4 3 4 3 T/R1 R2
TB/R Ri TN/R TB/R Ri TN/R SCPA SCPA
HYBRID HYBRID MCPA
1 2 1 2 T/R1 T/R2 T/R1 T/R2 T T1 T2 T1 T2 T
T/R1 T/R2 T/R1 T/R2 T/R1 R1 T/R2 R2 T/R1 T/R2 T/R1 T/R2 T/R1 T/R2 T/R1 T T1 T2 T3 T4 T T/R2
GSM-1900 UMTS-1900 GSM-1900 UMTS-1900 GSM-1900 UMTS-1900 GSM-1900
Figure 80. Hybrid Figure 81. Low Loss Figure 82. Multi- Figure 83. Single-
Combining Cable Combiner (LLC) with Carrier Power Carrier Power
Load Used to Ensure Integrated Duplexer. Amplifier (MCPA) with Amplifier (SCPA)
Low Passive RX Distribution integrated RX installed between
Intermodulation. Provided from GSM distribution. radios and bts
BTS. duplexer.
9.7.5. HYBRID COMBINING
Hybrid combiners provide a low cost method to simultaneously combine TX signals and divide RX signals. The main
disadvantage is the high loss incurred in both directions, which increases steeply with the number of ports combined.
When used with typical TX power levels, there is significant heat dissipation that must be managed. Therefore, this
104/135
solution has limited popularity in cell site situations and is rarely used to combine more than two ports. It is more
widely deployed for in-building coverage and similar applications.
9.7.6. LOW LOSS COMBINER—MULTIPLEXER
The Low Loss Combiner (LLC) is an alternate method to combine base station transmitters. Insertion loss in the LLC
is significantly lower than in a hybrid combiner and it is less expensive compared to power amplifiers. On the
downside, the LLC places constraints on the frequency ranges available to each system being combined.
Like a crossband coupler, an LLC is a filter multiplexer. While the former combines frequency bands spaced apart
fairly widely, the frequency ranges combined in an LLC are all in the same band, leading to very small gaps– guard
bands – between the ranges.
In some applications, guard band spectrum must be left unused and in order to minimize waste, the LLC is required to
accommodate the smallest guard band possible without compromising insertion loss or distorting the signals. Such
requirements lead to designs of increased size and cost. In other cases, the guard band spectrum can be reused on
a second feeder and antenna. Thus, the LLC specifications may be relaxed and more economical design options can
be chosen.
The LLC apportions a part of spectrum to each system. Future growth and de-growth, and other spectrum
redistributions will require reconfiguration, retuning, or replacement of the LLC. Remotely tunable equipment has
been introduced to partially alleviate these restrictions.
While LLCs are primarily employed for downlink combining, a corresponding filter multiplexer can also serve to
distribute uplink signals. It holds the advantage of not requiring supply power but is overall less popular than the
receiver multi-coupler method described below.
Some base stations on the market also have the ability to distribute part of the signal from each RX branch after the
receiver LNA. This makes it possible to design a smaller combiner unit without the need for other compensation of
uplink combiner loss.
9.7.7. SINGLE CARRIER POWER AMPLIFIERS (SCPA)
Preceding a hybrid combiner, an SCPA compensates for the combining loss. Being a non-linear amplifier, an SCPA
can itself be quite energy efficient but is suitable only for GSM and some similar technologies. Overall efficiency is
reduced by hybrid losses and need for heat management. The relatively simple amplifier circuits moderate unit cost.
105/135
ANT ANT TX
A’/A B/A’/B’ A’/A B/A’/B’

+
TXN RX RX TX TX
TXB/RX RXin TXN/RX TX/RXA TX/RXB TX1 TX2
Figure 84. LLC combines Figure 85. Filter multiplexer Figure 86. SCPA module
narrow portion of TX band into for downlink and uplink-- amplifies and combines two
broadband path. Includes quadriplexer. transmitters into one path.
duplexer for RX re-injection.
9.7.8. MULTI-CARRIER POWER AMPLIFIER (MCPA)
The MCPA is a high power linear amplifier capable of combining and amplifying multiple RF carriers into a single
output. An MCPA usually includes one or more amplifier “bricks” that work in parallel to provide the needed output
power. Input circuits can be expanded to accommodate from two up to eight or more ports. Integrated duplexers
route the uplink signals around the MCPA and RX amplification/distribution can be incorporated.
Advantages of the MCPA include the ability to boost downlink power for increased coverage and/or capacity as well
as complete frequency agility, placing no restrictions on what channel frequencies are used within a license band.
They can be used with any air interface technology, even combining multiple technologies. On the other hand, large
MCPA systems can entail a significant investment and they are not very energy efficient.
Modern base stations, such as for UMTS and LTE, mostly include MCPAs as an integrated part of the Radio Unit.
Here the Power output of the PA is closely monitored and adapted to current traffic situations as part of the overall
base station performance management. As an external MCPA will be connected outside of this monitor and control
loop it is less suitable for co-siting with this type of base stations.
106/135
TX/RX1 RX2
RX TX
+
RX TX TX RX
TX/RX1 TX3 TX4 TX/RX2
Figure 87. MCPA for one sector with two Figure 88. MCPA for three sectors, two duplexed
duplexed inputs and two simplex inputs, inputs and six simplex inputs per sector, one PA
three bricks installed. brick per sector, and one hot standby brick.
9.7.9. RECEIVER MULTICOUPLER
A Receiver Multi-coupler (RXMC) distributes uplink signals from shared antennas to multiple receivers. A power
splitter divides the input signal to a number of output ports. Preceding the power splitter, a Low Noise Amplifier (LNA)
compensates the splitting loss, preserving uplink sensitivity. At the input to the LNA, a preselector filter may be
integrated with the RXMC or exist as a part of other site equipment.
Characteristically, the RXMC distributes the full RX frequency band to all outputs. The net gain or loss is often the
same for all outputs but unequal division is also a design option where the RXMC serves receivers with differing input
level requirements or where the signal will be further divided before reaching the receivers.
For applications with few outputs – usually just two – a power splitter without LNA is sometimes used.
9.7.10. INTEGRATED DEVICES—SAME BAND COMBINER
The generic term Same Band Combiner (SBC) refers to a device that passes uplink and downlink signals,
incorporating some combination of same band transmit combining and/or receive signal distribution using any of the
methods described above, excluding power amplifiers. The configurations vary with application.
107/135
X X X
X X X
X X X
850 850 850
ANT ANT
DUPLEXER DUPLEXER
TX RX RX TX
ANT ANT IN IN A1 A2
QUADPLXR QUADPLXR RXMC RXMC SBC
T/RA T/RB T/RA T/RB 1 2 1 2 T/R1 R1 R2 T/R2
T/R1 T/R2 T/R1 T/R2 T R1 R2 R1 R2 T T/R1 R2 R1 T/R2

850-A”/A 850-B/A’/B’ GSM-850 UMTS-850 GSM-850 UMTS-850
Figure 89. Antenna Figure 90. RX distribution to Figure 91. Antenna sharing using
sharing using TX/RX simplex BTS using duplexers integrated SBC.
quadriplexer. and RXMC.
RXin
1:n
+
Figure 92. Dual RXMC with Eight Outputs per Cannel.
1 2 3 ··· n
108/135
ANT1 ANT2
TX RX RX TX
1:2 1:2
+ +
RX RX
TX/RX1 RX1 RX2 TX/RX2
Figure 93. Integrated Same Band Combiner (SBC).
9.7.11. TOWER MOUNTED AMPLIFIERS
A Tower Mounted Amplifier (TMA) is installed near the receiving antenna. Its purpose is to boost the uplink signal
before it is degraded by losses in the RF path to the receiver, thereby improving signal quality and base station
sensitivity.
9.7.12. BENEFITS OF TOWER MOUNTED AMPLIFIERS
TMAs provide benefits for all air interface technologies. Their fundamental effect on the uplink is an improvement in
carrier-to-noise ratio (C/N). Improvement of around 5-6 dB is typical but the amount varies according to RF path
configuration. This additional link budget margin can be utilized to improve various network performance parameters
such as:
• Coverage – in terms of cell radius as well as weak spots and indoor locations
• Accessibility – failed access attempts
• Retainability – dropped call rates
• Co-channel interference – call capacity in spread spectrum systems
• Data throughput – enabling higher order modulations
• Average handset output power – battery drain
109/135
Adjusting base station settings, the operator can optimize between the available benefits to obtain the best overall
Quality of Service (QoS).
9.7.13. CONFIGURATIONS
The key element in the TMA is the Low Noise Amplifier (LNA). It is always preceded by a preselector filter. Today’s
TMAs are universally of the dual duplex type, which allows use in duplexed feeders. Thus, two additional filters are
included to pass the downlink and isolate the LNA. Most often, the filter bandwidths cover a full license band but
other variants occur, including filters with additional out-of-band rejection to improve BTS interference immunity.
A TMA can be a single band device as above. Increasingly, single-band TMAs are provided as integrated pairs in
dual, or twin, assemblies in order to reduce cost, weight, and the number of devices on the tower.
Dual band TMAs combine two single band TMAs in one device. There may be a separate RF path for each band or
they may be diplexed into a single path at the BTS port and/or ANT port. Diplexers can also be integrated with single
band TMAs creating units with non-amplified RF paths often referred to as bypass or pass-through configurations.
Deployment of additional license bands is leading to further consolidation into TMAs supporting more than two RF
paths and frequency bands, again driven by efforts to avoid additional units of tower mounted equipment.
9.7.14. ENHANCED FEATURES—AISG
The Antenna Interface Standards Group (AISG) is an organization whose membership includes the majority of leading
wireless equipment manufacturers and many major wireless operators. Since its foundation in 2001, AISG has driven
the development of a protocol for communication between base stations and tower top equipment, including
antennas, TMAs, and other devices. The universally adopted AISG protocol facilitates control and monitoring of
functions such as antenna down tilt and TMA alarms.
Early implementations utilized separate cables for the connection between base and tower. More recently, a system
is increasingly favored where the signals are borne by a 2 MHz carrier on one or more feeder cables. The
components in the RF path must then be designed for compatibility with AISG communication whether they process
the signals or simply pass the low frequency carrier for example crossband couplers may have DC and control signal
bypass or block on one or more of their branches.
The AISG protocol provides a platform for implementation of advanced features in tower top equipment, allowing
better performance optimization and diagnostics while ensuring interoperability between any devices supporting the
standard. Adding remote connectivity to the AISG system will also reduce the frequency of site visits, having
beneficial impact on the operating budget.
110/135
Figure 94. Single band Figure 95. Twin single band TMA with Figure 96. TMA with
TMA. AISG support. integrated diplexer.
Figure 97. Dual band TMA with AISG support. Figure 98. Diplexed dual band TMA with AISG
support.
Figure 99. Dual diplexed dual band TMA with Figure 100. Diplexed dual band TMA with pass-
AISG. through and AISG.
111/135
9.8. MIMO CAPABLE DRIVE TEST EQUIPMENT
9.8.1. DRIVE TEST OVERVIEW
Users of wireless networks are by nature mobile. Whether riding in moving automobiles, walking through shopping
centers, running through airports, or traveling in mass transit vehicles, wireless customers expect their conversations
to be uninterrupted and their data devices to remain connected. Mobile test systems both drive and walk-test are
required in order to replicate the user experience when measuring network performance. These mobile test systems
record measurements tagged with time and position to zero on the exact locations of network problems and the time
of day those problems occur. Fixed location measurement systems simply are not up to the task.
The scope of drive testing may include optimizing and maintaining a service provider’s own network or may extend to
measuring and evaluating the competitor’s networks as well. As major carriers compete for customers by featuring
coverage maps as a major component of their advertising campaigns, an accurate picture of the coverage and
capabilities of all competing networks in a region has become an economic necessity.
Typical benchmark drive test systems consist of the measurement platform connected to UE devices such as phones
and wireless data modems assigned to the various service provider networks in a region as well as network
independent digital scanning receivers. The test devices are generally directly connected to the personal computer
running the measurement software by Universal Serial Bus (USB) cables for small to moderate size systems. For
larger benchmark systems, a hardware platform is utilized to connect and power the test devices and transfer data to
the measurement software.
The core mission of drive testing is to measure wireless network coverage and signal strength. Only drive-testing can
accurately measure signal strength as experienced by the wireless network subscribers as they travel along major
highways and secondary and secluded back roads. Digital scanners, capable of measuring and decoding the
technologies employed by the network as well as UE devices, generally measure signal strength.
Drive testing also evaluates the ability of subscribers to access the network whether originating voice calls, sending
SMS/MMS messages, or initiating packet data sessions to access the Internet or send emails. Since even strong RF
coverage is no guarantee that the network has the capacity to support voice and data calls in specific locations, drive-
test systems employ UE devices running automated scripts to place voice calls, and initiate data tasks while
measuring the success or failure of those tasks.
For voice calls, the call success rate, time required for call origination, and messaging between the UE and base
station is recorded as well as the ability of the network to maintain the calls.
For packet data sessions, the ability to successfully initiate a connection to the network and maintain that connection
is measured with particular emphasis on data throughput rates.
9.8.2. SCANNING RECEIVER MEASUREMENTS VERSUS USER EQUIPMENT (UE)
A new generation of sophisticated digital scanning receivers is now available for testing new technologies such as
LTE. These modular scanners are capable of simultaneously measuring multiple bands and decoding multiple
technologies using a single scanner. Plug in receiver modules extend the capabilities of these scanners to measure
new technologies and bands as spectrum is released for emerging technologies.
112/135
Scanning receiver provides:
• Unlike UE devices which are assigned to specific channels, scrambling codes, or PNs by the wireless
networks, digital scanners provide independent measurements of the wireless network. While the
scanners can be configured to follow the channel assignments of a UE device, the tasking of the
scanner is solely at the discretion of the drive-test technician.
• Digital Scanners have the ability to show individual components of the decoded signals. For
example, an LTE Cell ID scan provides measurements such as IPSS, PSS Io, SSS Io, Epss/Io,
Esss/Io, Composite Epss/Io, Composite Esss/Io, RSRP Port 0, RSRQ Port 0, RSRP Port 1, RSRQ
Port 1, Delay Spread, and Multipath Count for each Cell ID detected in a single scan.
• The Index of the Primary Synchronization Signal (IPSS) provides a relative or absolute time with
GPS. This can be used to ensure all cell sites are timed appropriately. The differential times
between cell sites can be indentified during drive testing to aid in the optimization for MIMO and
highest throughput performance. In addition, if any cell sites are not locked to GPS this index will drift
with time relative to the scanner, which is locked, to GPS.
• Some of the new generation scanners feature onboard Digital Signal Processing (DSP) and Field
Programmable Gate Arrays (FPGAs). This onboard digital processing and high-gate count
programmable logic extends the capabilities of the scanner to intelligently survey a network, detect,
locate, and process multi-technology signals rapidly. With the added capability to re-FLASH the
firmware in the field product life cycle of the scanner technology is increased significantly over
previous generations of scanners. The overall product cost is greatly reduced.
UE provides:
User equipment including phones, mobile data cards, and wireless USB modems, is required in a complete drive-test
system to fully test the subscriber experience.
• All wireless networks are affected by subscriber loading. To test voice call capacity, the drive-test
platform can task a number of UE phones to simultaneously place calls to stress the network’s ability
to handle the load simulating performance during rush hour traffic, large sports or entertainment
events, or emergencies. To test data capacity, a number of UE data modems can be tasked to
download/upload files, access Web pages, and send emails to ensure that the network can handle
the data transfer loading.
• The drive-test platform records the success and failures of the UE devices as they perform their
assigned voice and data tasks along with time and location information into a log file. For voice calls,
statistics such as failed originations, dropped calls, and no service failures, and successful call starts
are captured. For data sessions, statistics such as successful connects, failed connects, no data
service, dropped connections, and data throughput are captured.
• UE devices can also be captured in the log file of the drive-test measurement platform. Layer 1 and
layer 3 communications between the BTS and phone are captured and recorded in the log file. For
data sessions, TCP/IP communications are captured and can be viewed in a protocol analyzer
application. This communication dialog can be instrumental in tracking down and correcting network
issues.
113/135
9.8.3. IMPACT OF NEW TECHNOLOGIES ON DRIVE TESTING SCANNERS
The higher data rate capabilities in the cellular environment naturally require higher bandwidths and higher
performance in drive test-technology. Scanners capable of sampling IF signal bandwidths up to 20 MHz are required
for the new generation of technologies including LTE. With the addition of MIMO, it is desirable for scanners to also
provide a two-channel MIMO capability while beam steering effectiveness will require up to four channels.
Support for multi-antenna techniques
• Receive diversity at the mobile device. One of the simplest and conventional diversity methods, a
form of multi-antenna implementations, is the use of two antennas at the receiver end. Based on
RSSI, the best antenna source is selected for signal reception. While this technique may improve
reception under certain conditions where signal-to-noise (SNR) ratio suffers due to multipath fading,
data throughput cannot be improved without using other techniques.
• Transmit diversity using space/frequency block coding (SFBC) at the base station Transmit diversity
is another conventional method and defined in LTE specifications as evolved Node B, or eNB). In
LTE, Space Frequency Block Coding (SFBC) is used to improve signal robustness under fading
conditions. The transmitters send the same underlying user data, but in different parts of the RF
frequency space.
• MIMO spatial multiplexing at the eNB, for one or two users. True MIMO, with two transmitters and
two receivers with independent data content, is also known as spatial multiplexing. Each receiver
sees the output of the channel, which is a combination of the outputs from the transmitters. Using
channel estimation techniques, the receivers use matrix mathematics to separate the two data
streams and demodulate the data. In ideal conditions, data capacity would be doubled, though there
is a premium to be paid in better SNR requirements than for SISO. Practically, the doubling of data
capacity is never achieved, but definite increases in data capacity can be seen.
• Cyclic Delay Diversity (CDD) at the eNB, used in conjunction with spatial multiplexing. CCD
introduces deliberate delays between the antennas to create artificial multipath. It is applied more
dynamically in LTE than in other radio systems.
• Conventional phased-array beam steering introduces phase and amplitude offsets to the whole of the
signal feeding each transmitting antenna. The basic idea is to focus the signal power in a particular
direction. The same technique of applying phase and amplitude offsets can be used on the receiving
antennas to make the receiver more sensitive to signals coming from a particular direction. In LTE,
the amplitude and phase of individual resource blocks can be adjusted, making beam steering far
more flexible and user-specific. Although beam steering does not increase data rates, the end result
is similar to diversity in terms of increasing signal robustness. The effectiveness of beam steering
increases with the number of transmitting antennas, which allows for the creation of a narrower beam.
The gains possible with only two antennas are generally not considered worthwhile; thus, beam
steering is generally only considered for the four-antenna option. Scanner technology, which is
capable of up to 4 receiver antenna signal paths, can be used to identify the effectiveness of beam
steering.
The advancements in multi-antenna technologies to achieve greater capacity and higher data throughput in cellular
technologies necessarily demands new scanner technology, which can provide dynamic real-time measurements of
the effectiveness of the various antenna and processing techniques. A real-time, channel capacity measurement
114/135
capability would be especially valuable. Advances in scanner technology is expected in response to the need
operators will have to evaluate and optimize the performance of their next generation networks.
9.9. INDOOR DISTRIBUTED ANTENNA SYSTEM—MIMO COVERAGE
In many countries, more than 70 percent of the traffic on cellular networks originates from or terminates inside
buildings. Driven by an increasing demand for mobile data services, the iPhone effect, this number is expected to rise
32
even higher within the next years. When rolling out new systems addressing high-speed data needs, e.g., LTE or
HSPA+, a cost effective solution is explored by extending the outdoor macro coverage to inside. In order to reduce
the building penetration path loss a relay can be used to amplify and forward (A&F) or decode and forward (D&F) the
signal. The benefits of a D&F relay on the achievable LTE downlink indoor coverage extension has been
33
demonstrated in “Relaying in Long Term Evolution: Indoor Full Frequency Reuse.” Implementing the D&F relay,
however, adds complexity to the system and decreases overall air-interface capacity due to the in-band backhauling.
In the A&F mode, the relay terminal simply amplifies and retransmits the signal received from the source terminal.
Figure 101 compares the results measured at the Heinrich Hertz Institute in Berlin by applying both relay
technologies.
Data Throughput
A) B)
Figure 101. Graphical Representation of the in-building coverage measured with: A) A D&F relay where 50
32
percent of the link capacity is allocated for in-band backhauling, And B) An A&F optical DAS with an
34
interleaved antenna arrangement.
34
In order to obtain the best cost-to-coverage solution several antenna configurations have been investigated. Due to
the fact that an interleaved distributed antenna arrangement can be set-up by re-cabling of an existing single antenna
2G or 3G distribution system this solution turned out to be the most advantageous. At a minimal upgrade cost, it
provides exceptional data rates and the LTE pre-coding supports the transmission of two separate spatial data
streams at two different antennas. It is therefore possible to fully exploit the rich scattering that exists in most in-
building environments.
In addition to the in-building coverage test, it could be shown that a MIMO coded signal can be transmitted without
any performance degradation in good LOS conditions. This is good to remember when the distribution system has to
be fed by picking up the signal off air.
In the face of all the enthusiasm about the possibility to boost the in-building data rate by deploying MIMO
technologies, it has to be mentioned that a sufficient Signal-to-Interference-and-Noise Ratio (SINR) is the price, which
115/135
has to be paid. As shown in Figure 102 A) a minimum SINR of 26 dB is required to enhance the capacity of a single
input single output (SISO) system by 50 percent. To justify the cost for a second RF infrastructure it is assumed that
a data rate enhancement of 70 to 80 percent (i.e., a SINR of 40 dB to 50 dB is required). For these capacity
35
calculations, a measured average channel condition number of 14 dB, valid for a small/office in-door environment,
has been assumed. To estimate the required RF power per antenna at 2.6 GHz in a typical indoor environment the
36
scenario B3 (hotspot large indoor hall) based on the WINNER propagation model is shown in Figure 102. Hence
depending on the number and thickness of the walls to be taken into account, a similar path loss can be expected for
scenario A1 (indoor office / residential). For simplicity, let us assume a minimum UE receiver’s sensitivity of -100
dBm in a 20 MHz bandwidth. Taking into account 60 to 80 dB path loss, a RF power level of -40 to -20 dBm at the
receiver is required. Targeting for a minimum data rate enhancement of 70 percent this results in 0 to +20 dBm
radiated power at each antenna per carrier bandwidth. Alternatively, to increasing the RF power, the number of
antennas can be increased and thus the spacing in between reduced. It also should be mentioned that the situation
can be improved by operating at lower frequencies, e.g., at 700 to 800 MHz where the path loss is expected to be 10
to 15 dB lower than at PCS frequencies.
A) B)
Figure 102. Graphs for MIMO link budget considerations: A) data rate enhancement MIMO vs. SISO as a
function of SINR, and B) typical indoor path loss graphics @ 2.6 GHz (WINNER II channel model).
Another point of interested, especially for in-building MIMO distribution systems, is the minimum spacing in-between
the antenna dipoles to ensure the maximum data rate enhancement. The amount of additional transport capacity,
provided by different spatial MIMO streams, is directly linked to the de-correlation of these streams. Since the
correlation between two antenna dipoles is increasing the closer they come it might be expected that the data rate
enhancement may drop accordingly. However it has been shown that local scatterers, in the near vicinity of the
37
antenna elements, significantly decreases this effect. It turned out that not the antenna spacing but the angular
spreading of the incident waves is the determining factor for the antenna de-correlation. Especially in rich scattering
indoor environments it has been proven by experiment that a spacing of λ/2 is already enough to guarantee a
38
reasonable data rate enhancement.
In low scattering indoor environments, e.g., large hallways or conference rooms, a low angular spreading has to be
39
expected. A phenomenon, which is well known from outdoor, above rooftop, antenna installation. Unless large
antenna spacing, e.g., interleaved indoor antenna arrangements (as described in the previous paragraph) can be
used, it is assumed that cross-polarized antennas are advantageous for indoor installations as well. Results show
that increasing the de-correlation by polarization performs better than increasing the antenna spacing in strong line of
40
sight (LOS) scenarios. In non line of sight (NLOS) scenarios, the results are of the same order.
116/135
10. TERMINAL ANTENNA ISSUES
10.1. PROSPECTS AND CHARACTERISTICS OF MULTIPLE ANTENNAS IN TERMINALS

3
In a previous 3G Americas white paper on MIMO and smart antenna technology, many of the problems faced by user
equipment (UE) vendors were discussed. The problems range from issues of antenna correlation and coupling, the
challenges of designing and manufacturing small-sized UE devices, multi-band support and user effects. General
knowledge of the solutions to these problems is fairly well known. For instance, as antennas are separated, they tend
to have less correlation and coupling. It is very hard to generate specific data on the impact of real antennas on smart
antenna systems because the implementations will be diverse across device size and designer preferences (antenna
type, polarization and/or pattern diversity). Accumulation of this data will have to be done with the large-scale
deployment with many devices. Carriers are probably in the best position to catalogue this information given the large
number of deployed devices.
In order to provide insight into the effects that actual deployment of UE antennas will have on smart phones sized
devices, we provide an analysis that attempts to address the two most important parameters in wireless
communication systems: link budget and capacity impact. One of the first things that a system designer is going to
want to know is the antenna gain of the UE, given that there are two antennas now required for LTE. In other words,
how much additional gain can be expected with the use of both antennas relative to a single antenna. Given that
techniques such as pattern and polarization diversity can be employed to help de-correlate arriving signal paths, the
question arises as to what the impact of the antenna patterns is going to be.
41
To investigate the link budget impact, a study was conducted using an approach similar to that described in. This
study was done with antenna types that are likely to be used within a typical smart phone dimension for use in LTE
MIMO mobile systems. The objective of the study was to try to characterize the relative performance improvement in
the link budget that would be realized in a MIMO-based LTE system from the use of two receive antennas relative to
the single antenna systems most commonly used in current UEs.
This study was set up with the following two scenarios.
1. A phone placed in between a phantom head on one side and a phantom hand placed on the opposite side of
a phone. This provides a realistic situation in which to examine antenna performance for mobile devices
since this is a very common usage pattern of users. This scenario provides fairly large absorption and
shielding losses because the antenna is shielded by both the head and hand. However, even though the
mutual coupling between the antennas tends to increase the magnitude and phase of the radiation pattern
changes such that the antennas tend to be less correlated than in a free space scenario. This de-correlation
tends to overcome the absorption loss and increases the performance relative to two antennas in free space.
This scenario is illustrated in Figure 103 and is referred to as the Voice Mode.
2. A phone placed in front of the body held with the hands. This is a common usage scenario for texting or web
surfing. This scenario is illustrated in Figure 103 and is referred to as the Data Mode
117/135
(a) Voice Mode (b) Data Mode
Figure 103. Position of the handheld device for the two scenarios.
In our study, we used several different head and hand models and averaged the results. This probably accounts for
41
most of the differences in performance between our study and reference. However, the exact form factor, antenna
location, antenna type and orientation are also factors that will yield different performance values. The study was
42 43
carried out via software simulations with the GEMS and FEKO software tools as well as anechoic chamber
measurements, which confirmed the simulations given the close agreement between the two sets of data. Table 5
shows the results for the Voice Mode. Table 6 shows the results for the Data Mode. The “Improvement” column is
the data that was simulated and measured in this study. The “Improvement from” column is the data found
41
referenced in, “Actual Diversity Performance of a Multiband Diversity Antenna with Hand and Head Effects.” Note
that the simulated and measured results include diversity gain, and are relevant to control channel reception in a UE.
Table 5. Link budget receive performance improvement due to two receive antennas over one receive
antenna on a handheld placed between a phantom head and hand.
Frequency (MHz) Improvement (dB) Improvement [4] (dB)
700 2 to 5 --
850 2 to 5 5-6
950 3 to 6 5 to 6
2100 6 to 9 8 to 10
2600 6 to 9 --
118/135
Table 6. Link budget receive performance improvement due to two receive antennas over one receive
antenna on a handheld placed in front of user in a typical data usage fashion.
Frequency (MHz) Improvement (dB) Improvement [4] (dB)
700 4 to 6 --
850 5 to 7 8 to 10
950 5 to 8 8 to 10
2100 7 to 10 8 to 10
2600 7 to 10 --
To get an indication of the possible loss of capacity that may occur when using two realistic antennas, we calculate
the maximum capacity possible under ideal conditions using uncorrelated isotropic antennas and compare those
results to that of a design that would contain realistic antennas. The realistic design for a smart phone will utilize the
well-known inverted-F antenna (IFA) technology. The design is cross-polarized and we look at two frequencies: 2.6
GHz (190 MHz bandwidth) and 780 MHz (40 MHz bandwidth). These are two prominent LTE frequency bands. In this
study, we do not take into account the effects of the head and hands.
To measure the maximum capacity we use Shannon’s generalized capacity equation
ρHH H
C = log 2 det( I N + )
NT
where ρ is the average signal-to-noise ratio, H is the channel at each subcarrier, IN is the identity matrix and NT is the
number of transmit antennas. With this equation, if we can impose the effects of the antennas on the channel matrix,
we can generate the upper limit of the capacity of a single MIMO radio link with equal power transmitted per layer.
44
This can be done by importing the antenna patterns into the spatial channel model (SCM that is used extensively by
3GPP) and generating H with their effects. The parameters for 780 MHz that are plugged into the model are shown in
Table 7. As you can see we assume that the base station antenna configuration can be modeled as two dipoles
separated by 20 wavelengths (i.e., they are uncorrelated). We focus on the Urban Macrocell channel. Figure 107 and
Figure 108 exemplify transmit and receive antenna configurations that contain angle spread and orientation.
119/135
Table 7. SCM parameters for 2X2 780 MHz antenna system.
SCM Urban Macrocell Parameters BS MS
Azimuth Angle Spread 8 deg 68 deg
Azimuth Array Orientation 0 deg 0 to 180 deg
Elevation Angles 0 deg -90 to 90 deg
Number of Antennas 2 2
Element Separation 20 λ 0.12 λ (cross-polarized)
# Independent Channel Realizations 10,000
Figure 105. Channel parameters.

Figure 104. MIMO channel model.
In an effort to isolate the effect of realistic antennas, we compute the capacity at a given average SNR at the antenna
input by averaging the Shannon capacity over multiple independently faded channel realizations. By further
45
accounting for the angular power distributions in azimuth and elevation, we can compare the results of a realistic set
of antennas with the results of doing the same with an isotropic antenna and gain insight into the capacity impact of
more realistic antennas.
Figure 106 shows the result of this for both 2.6 GHz and 780 MHz antennas. Note that the orientations are averaged
to provide one resulting curve. UE is moving at the speed of 10 m/s. Looking at the 10 dB SNR mark, it can be seen
that an extra dB of SNR for the UE with an IFA antenna pair is needed to achieve the capacity upper limit that you
would get with an isotropic antenna pair. At 10 dB SNR, there is a loss of 0.5 bps/Hz or 9.5 percent for the UE with
2.6 GHz IFA antenna pair. For 780 MHz antenna pair, there is a loss of 1.8 bps/Hz or 35 percent at 10 dB. Note that
in all the simulations, the effect of antenna efficiency is incorporated.
120/135
Figure 106. Shannon upper limit capacity at 2.6 GHz and 780 MHz in an urban macro cell environment with
two transmit and two receive antennas.
In conclusion, we have examined the effects of realistic antennas in a smart-phone form factor on link budget
performance and maximum link capacity. We have noted from a study performed for this document that there will be
a link budget improvement of 2 dB at lower frequencies to around 6 dB at some of the higher LTE frequency bands.
Others have shown greater gains than that. On the other hand, we see that realistic antenna designs in realistic
environments will see a loss of maximum capacity potential of over 30 percent.
10.2. CURRENT STATUS
Multiple antennas are currently found in the terminals of many wireless products, particularly PCs, data cards and
USB dongles. See for example, the products whose antennas are photographed below in Figure 108 through Figure
110 below.
It may be instructive to recall that terminals are not just handsets anymore. Mobile terminals such as the Kindle and
iPad, not to mention laptops and notebooks that have integrated UT built in have a great deal of room for multiple
antennas. Beyond the FM radio antenna, GPS receiver, WiFi and multiple band antennas, they also now support
multiple antennas as well. In these emerging pad computers and readers, we see sufficient room to support more
than 2 antennas, though they all appear to use two at this time. See for example the two PCS band antennas used in
the Kindle reader in Figure 111.
121/135
Figure 107. EVDO data card clearly showing two-branch diversity.
122/135
Retractable Antenna
Internal Antenna
Figure 108. Photograph of the inside of the Pantech UM150 EVDO USB dongle with 2 selectable
antennas.
123/135
GSM antenna
900, 1800 and Multi-band antenna for Treo 600
1900MHz
UMTS
Samsung SCH-V740 ("The Blade")
Figure 109. Example antennas from representative handsets, illustrating manufacturing and design
approaches.
A commercial WiMAX terminal (Cmotech CMU300) with receive diversity is revealed in the figure below:
124/135
Figure 110. Internal view of the commercial CMU 300 modem with receive diversity. The two
diversity antenna ports are indicated.
125/135
Figure 111. Photo of the inside of the Amazon Kindle showing two antennas for PCS band use.
126/135
11. SUMMARY AND CONCLUSIONS
A variety of smart antennas, from those with remote controlled boresight to beam-steering arrays along with
beamforming MIMO as well as active antenna array antenna configurations have been reviewed in great depth. Their
installation and operational characteristics have been detailed and several deployment and upgrade strategies have
been discussed.
General conclusions about how best to upgrade a network are not easily made because every operator has a unique
set of constraints such as what frequencies are available, what air interfaces preexist and what network, antenna and
user equipment has previously been deployed. Even so, the wide variety of antenna types and processing techniques
makes impressive improvements in the capacity and coverage of emerging 4G networks.
127/135
12. ACKNOWLEGEMENTS
The mission of 3G Americas is to promote, facilitate and advocate for the deployment and adoption of the GSM family
of technologies including LTE throughout the Americas. 3G Americas’ Board of Governor members include Alcatel-
Lucent, Andrew Solutions, America Móvil, AT&T (USA), Cable & Wireless (West Indies), Ericsson, Gemalto, HP,
Huawei, Motorola, Nokia Siemens Networks, Openwave, Research In Motion (RIM), Rogers Wireless (Canada), T-
Mobile USA, and Telefónica.
We would like to recognize the significant project leadership and important contributions of Stephen A. Wilkus of
Alcatel-Lucent and Kevin Linehan of Andrew Solutions as well as the other member companies from 3G Americas’
Board of Governors who participated and contributed to the development of this excellent white paper.
128/135
13. ACRONYMS
3G Third Generation
3GPP Third Generation Partnership Project
AA Adaptive Array
ABF Adaptive Beamforming
AFLA Another Four Letter Acronym
AISG Antenna Interface Standards Group
ARLBR Active Radio Link Bit Rate
AWS Advanced Wireless Service
BF Beamforming
BRS Broadband Radio Service
BS Base Station
BSA Base Station Antenna
CAPEX Capital Expenditures
CDD Cyclic Delay Diversity
CDF Cumulative Distribution Function
CDMA Code Division Multiple Access
C/I Carrier to Interference Radio
CL Closed-Loop
CPRI Common Public Radio Interface
CQI Channel Quality Indicator
DL Downlink
DPD Digital Pre-Distortion
DSP Digital Signal Processing
EBS Educational Broadband Services
Ec/No Signal Energy per Chip over Noise Power Spectral Density
EDGE Enhanced Data Rates for GSM Evolution
129/135
eNB Evolved Node B
EPC Evolved Packet Core
ePDSN Evolved PDSN
EPS Evolved Packet System or Service
ERP Effective Radiated Power
e-UTRAN Evolved UMTS Terrestrial RAN
EVDO Evolution-Date Optimized
FDD Frequency Division Duplexing
FPGA Field Programmable Gate Array
GPRS General Packet Radio Service
GSM Global System for Mobile Communications
HSDPA High Speed Downlink Packet Access
HSPA High Speed Packet Access
IF Intermediate Frequency
IFA Inverted-F Antenna
IPSS Index of the Primary Synchronization Signal
LAN Local Area Network
LLC Low Loss Combiner
LNA Low Noise Amplifier
LTE Long Term Evolution
MCCH Multicast Control Channel
MCPA Multi-Carrier Power Amplifier
MIMO Multiple-Input Multiple-Output
MRxD Mobile Receive Diversity
MS Mobile Station
MU-MIMO Multiple User MIMO
OBSAI Open Base Station Standard Initiative
OFDM Orthogonal Frequency Division Multiplexing
130/135
OL Open Loop
OPEX Operating Expenditures
PA Power Amplifier
PCs Personal Computers
PCFICH Physical Control Format Indicator Channel
PCS Personal Communication Systems
PDCCH Physical Downlink Control Channel
PHICH Physical Hybrid-ARQ Indicator Channel
PMI Precoding Matrix Indication
QAM Quadrature Amplitude Modulation
QoS Quality of Service
RAB Remote Antenna Beam Width
RAS Remote Antenna Azimuth
RET Remote Electrical Downtilt
RF Radio Frequency
RI Rank Indication
RRH Remote Radio Head
RRU Remote Radio Unit
RU Radio Unit
RX Receiver
RXMC Receiver Multicoupler
SAW Surface Acoustic Wave
SBC Same Band Combiner
SCM Spatial Channel Model
SCPA Single-Carrier Power Amplifier
SFBC Space-Frequency Block Coding
SIC Successive Interference Cancellation
SIMO Single-Input and Multiple-Output
131/135
SINR Signal-to-Interference-Plus-Noise Ratio
SISO Single-Input and Single-Output
SNR Signal-to-Noise Ratio
SON Self-Organizing Network
SPR Sector Power Ratio
SU-MIMO Single User MIMO
TDD Time Division Duplexing
TD-SCDMA Time Division Synchronous CDMA
TMA Tower Mounted Amplifier
TRDU Transmitter-Receiver Duplexer Units
TX Transmit
UE User Equipment
UL Uplink
UMTS Universal Mobile Telephone System
USB Universal Serial Bus
UTRA Universal Terrestrial Radio Access
V-pol Vertically polarized
WIMAX Worldwide Interoperability for Microwave Access
X-pol Cross polarized (Dual Polarized)
132/135
14. ENDNOTES
1
Chetan Sharma Consulting, “US Wireless Data Market Q4 2009 and 2009 Update,” available on the web at:
http://www.chetansharma.com/US%20Wireless%20Market%20Q4%202009%20Update%20-%20Mar%202010%20-
%20Chetan%20Sharma%20Consulting.ppt#261,1 on March 3, 2010.
2
Information Content estimate from www.wolframalpha.com, March 3, 2010.
3
“MIMO Transmission Schemes for LTE and HSPA Networks,” 3G Americas Whitepaper, June 2009.
4
“Vocabulary of 3GPP Specifications,” TR21-905, 3GPP, version 8.8.0, March 2009.
5
Rysavy Research, “HSPAto LTE-Advanced: 3GPP Broadband Evolution to IMT-Advanced (4G),” 3G Americas
White Paper, September 2009.
6
Jonas Karlsson and Mathias Riback, “Initial Field Performance Measurements of LTE,” Ericsson Review No. 3,
2008.
7
Krishna Balachandran, Qi Bi, Ashok Rudrapatna, James Seymour, Soni, and Andreas Weber, “Performance
Assessment of Next-Generation Wireless Mobile Systems,” Bell Labs Technical Journal 2009 vol. 13:4, pages 35-58,
http://www3.interscience.wiley.com/journal/122211215/abstract last accessed on March 30, 2010.
8
A.M.D. Turkmani, A.A. Arowojolu, P.A. Jefford, and C.J. Kellett, “An Experimental Evaluation of the Performance of
Two-Branch Space and Polarization Diversity Schemes at 1800 MHz,” IEEE 0018-9545/94404,00, 1995.
9
Brian S. Collins, “The Effect of Imperfect Antenna Cross-Polar Performance on the Diversity Gain of a Polarization-
Diversity Receiving System,” Microwave Journal, April 2000.
10
R. Bhagavatula, R. W. Heath, Jr., and K. Linehan, “Performance Evaluation of MIMO Base Station Antenna
Designs," Antenna Systems and Technology Magazine, vol. 11, no. 6, pp. 14 -17, Nov/Dec. 2008.
11
Lawrence M. Drabeck, Michael J. Flanagan, Jayanthi Srinivasan, William M. MacDonald, Georg Hampel, and
Alvaro Diaz, “Network Optimization Trials of a Vendor-Independent Methodology Using the Ocelot™ Tool,” Bell Labs
Technical Journal 9(4), 49–66 (2005) © 2005
12
“Final Report on Semi-Smart Antenna Technology Project,” Ofcom Contract No. 830000081, Document No.
830000081/04, July 2006.
13
M. Petersen, et al., “Automatic Antenna Tilt Control for Capacity Enhancements in UMTS FDD,” IEEE 2004, 0-
7803-8521-7/04, 2004.
14
Mano D. Judd, Thomas D. Monte, Donald G. Jackson, and Greg S. Maca, “Tranceive Distributed Antenna
Systems,” U.S. Patent No. US 6,597,325 B2, July 22, 2003.
15
“Active Antenna Arrays, Small-Footprint, Scalable RF Solutions for Base Stations,” Bell Labs Research Project,
Alcatel-Lucent Innovation Days, December 2008. Available on line at http://innovationdays.alcatel-
lucent.com/2008/documents/Active%20Antenna%20Array.pdf.
133/135
16
“Nokia Siemens Network Networks and Vodafone Demonstrate Active Antenna Solution,” Press Release, July 29,
2009. Available on line at http://www.nokiasiemensnetworks.com/news-and-events/press-room/press-releases/nokia-
siemens-networks-and-vodafone-demonstrate-active-ant.
17
Osman N.C. Yilmaz, Seppo Hamalainen, and Jyri Hamalainen, “System Level Analysis of Vertical Sectorization for
3GPP LTE,” Proceedings of the 6th International Conference on Symposium on Wireless Communication Systems,
Siena, Italy, pp. 453-457, 2009.
18
Alan J. Fenn, Donald H. Temme, William P. Delaney, and William E. Courtney, “The Development of Phased-Array
Radar Technology,” Lincoln Laboratory Journal, Vol 12, no 2, 2000, page 321-340. Available on-line at:
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.73.2849&rep=rep1&type=pdf , last accessed April 17, 2010.
19
F. Athley, M. Alm, O. Kaspersson, K. Werner, J. Furuskog, and B. Hagerman, “Dual-Polarized Base Station
Antenna Configurations for LTE,” to appear in Proc. 2010 IEEE International Symposium on Antennas and
Propagation.
20
K. Werner, J. Furuskog, M. Riback, and B. Hagerman, “Antenna Configurations for 4x4 MIMO in LTE—Field
Measurements,” to appear in Proc. IEEE VTC Spring 2010.
21
“Spatial Channel Model for Multiple Input Multiple Output (MIMO) Simulations,” 3GPP TR 25.996, V7.0.0, available
on line at http://www.3gpp.org/FTP/Specs/html-info/25996.htm.
22
L. Wan, S. Tsai, and M. Almgren, “A Fading-Insensitive Performance Metric for a Unified Link Quality Model,” Proc.
IEEE WCNC, 2006.
23
“Physical Layer Aspects for Evolved Universal Terrestrial Radio Access (UTRA),” 3GPP TR 25.814, V7.1.0, section
A.2, available on line at http://www.3gpp.org/FTP/Specs/html-info/25814.htm.
24
Section 7.1 of 3GPP Section 7.1 of TS 36.101 V8.7 states: "The requirements in Section 7 assume that the
receiver is equipped with two Rx port as a baseline. Requirements for 4 ports are FFS. With the exception of clause
7.9 all requirements shall be verified by using both (all) antenna ports simultaneously." Available on line at:
http://www.3gpp.org/ftp/Specs/html-info/36101.htm last accessed on April 18, 2010.
25
City Code of Lindon City, UT, Title 5: Licensing and Francising, 5.07 Personal Wireless Telecommunications
Facilities, available on-line at: http://www.lindoncity.org/government.wireless.html, last accessed March 25, 2010.
26
“Illinois State Toll Highway Authority Telecommunications Facility Leasing Program” Illinois Tollway, October 2005,
available on-line at:
http://www.illinoistollway.com/pls/portal/docs/PAGE/TW_CONTENT_REPOSITORY/TW_CR_DOING_BUSINESS/TW
_CR_DOING_BUSINESS_LEASING/TOWERPROGRAM.PDF last accessed March 25, 2010.
27
“The Impacts of Antenna Azimuth and Tilt Installation Accuracy on UMTS Network Performance,” Bechtel
Telecommunications Technical Journal, Vol. 4, No. 1, January 2006.
28
http://www.bechtel.com/assets/files/TechJournal/2009/Comms%2003%20Performance%20Engineering%20Advances
%20to%20Installation.pdf
29
“Frequently Asked Questions for Passive IM Products,” Summitek Instruments,
134/135
http://www.summitekinstruments.com/passive/faqpim.asp.
30
Multi-Band PIM Calculator, Andrew Solutions, Software Tools,
http://www.commscope.com/andrew/eng/product/antennas/bsa/sw_tool/1205324_13426.html.
31
“Intermodulation Products of LTE and 2G Signals in Multitechnology RF Paths,” December 2009,
http://www.bechtel.com/assets/files/TechJournal/2009/Comms%2001%20Intermodulation%20Products%20of%20LTE
%20and%202G.pdf.
32
“In-Building Wireless Systems – Passive and Active DAS, Repeaters, Picocells, and Femtocells,” 2009
ABIresearch, www.abiresearch.com.
33
V. Venkatkumar, T. Wirth, T. Haustein, and E. Schulz, ”Relaying in Long Term Evolution: Indoor Full Frequency
Reuse,“ in European Wireless 2009, Aarlborg, Denmark, 2009.
34
T. Wirth, O. Braz, T. Haustein, J. Stefanik, and Lars Thiele, “LTE Amplify and Forward Relay for Indoor Coverage
Extensions,“ VTC 2010-Fall Ottawa, to be published.
35
Peter Cain, “Using Channel Condition Number to Visualize MIMO Performance,” IWPC MIMO, 12 November 2008.
36
IST-4-027756 WINNER, D1.1.2 V1.1, WINNER II Channel Models, 2007, Source: http://www.ist-winner.org.
37
W. Lee, “Effects on Correlation between Two Mobile Radio Base Stations,” IEEE Transactions on Communications,
21(11): pp. 1214-1224, 1973.
38
S. Caban, , C. Hehlfuhrer, L. W. Mayer, and M. Rupp, “2008 – 2x2 MIMO at Variable Antenna Distance,” IEEE
Vehicular Technology Conference, VTC Spring 2008, http://publik.tuwien.ac.at/files/PubDat_167444.pdf and
http://publik.tuwien.ac.at/files/pub-et_12543.pdf.
39
W. Lee and Robert H. Brandt, “The Elevation Angle of Mobile Radio Signal Arrival,“ IEEE Trans. Communications,
Vol. COM-21, pp. 1194-1197, November 1973.
40
J. Jiang, et al., “Correlation Analysis of Multiple-Input Multiple-Output Channels with Cross-Polarized Antennas,”
Proceedings of APCC2008, 2008.
41
“Actual Diversity Performance of a Multiband Diversity Antenna with Hand and Head Effects,” IEEE Transactions on
Antennas and Propagation, Vol. 57, No. 5, May 2009.
42
http://www.2comu.com/
43
http://www.feko.info/
44
http://www.ist-winner.org/3gpp_scm.html
45
“Angular Power Distribution and Mean Effective Gain of Mobile Antenna in Different Propagation Environments,”
IEEE Transactions on Technology, Vol. 51, pp. 823-838, September 2002.
135/135

Mimo and Smart Antennas For 3g and 4g Wireless Systems May 2010 Final

Încărcat de

Informații document

Descriere originală:

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Mimo and Smart Antennas For 3g and 4g Wireless Systems May 2010 Final

Încărcat de

Drepturi de autor:

Formate disponibile

1/135

1. BACKGROUND AND SCOPE OF WHITEPAPER ........................................................................................... 14

Figure 4. Illustration of the terminology of base stations and cells. 17

Figure 5. Early base station antenna technology. 19

Figure 6. Current base station antenna technology -- extensive use. 20

Figure 7. Latest base station antenna trends -- moderate usage. 21

Figure 8. Advanced Antenna Technology. 22

Figure 9. MIMO Systems. 23

Figure 10. Reconfigurable beam antenna. 24

Figure 11. Active Antenna System (AAS). 25

Figure 12. Typical multi-column planar array architecture. 29

Figure 13. Beamforming antenna and fixed beam quad-port antenna. 30

Figure 14. Topology of a single column of radiating elements. 31

Figure 15. Connector ports for vertical and dual-polarized arrays. 32

Figure 16. Typical antenna patterns for a single column of elements. 33

Figure 18. Azimuth Pattern Beamforming. 35

Figure 19. Azimuth pattern beamforming with a scan angle of 20 degrees. 36

Figure 22. Beamforming degradation due to amplitude and phase errors. 39

Figure 25. Fixed twin-beam antenna architectures. 42

Figure 27. Life-cycle of a wireless network. 45

Figure 31. Predictable changes in traffic distribution. 51

Figure 33. Radio Integration Trend. 54

Figure 34. 900 MHz Active Antenna. 54

Figure 35. Typical active antenna architecture. 55

Figure 36. Electronic tilt applications. 56

Figure 37. Self-healing capacity. 56

Figure 44. Tower top “pinwheel” with dimensions. 66

Figure 45. Common monopole during installation of antennas. 67

Figure 49. Antenna mount with remote radio head. 71

Figure 50. Angle below horizon for various tower heights. 74

Figure 53. Differences in the horizontal pattern cuts. 75

Figure 55. Combination of tilts. 77

Figure 62. Graphical representation of Sector Power Ratio (SPR). 83

Figure 65. Front-to-back and Cross-pol ratio. 85

Figure 67. Azimuth Adjustment Tool. 88

Figure 68. Block diagram of a RRH/TRDU. 91

Figure 73. Breakout and Termination Scheme. 99

Figure 74. Accessories. 100

Figure 75. Multiband sector with separate feeders. 102

Figure 76. Shared feeders using diplexers (crossband couplers). 102

Figure 78. Diplexer. 103

Figure 79. Triplexer. 103

Figure 85. Filter multiplexer for downlink and uplink--quadriplexer. 106

Figure 89. Antenna sharing using TX/RX quadriplexer. 108

Figure 91. Antenna sharing using integrated SBC. 108

Figure 93. Integrated Same Band Combiner (SBC). 109

Figure 94. Single band TMA. 111

Figure 96. TMA with integrated diplexer. 111

Figure 97. Dual band TMA with AISG support. 111

Figure 104. MIMO channel model. 120

Figure 105. Channel parameters. 120

Table 1. Antenna Types Supplied - One Vendor........................................................................................................... 25

Table 3. Number of Antennas per Sector. ..................................................................................................................... 65

Single Stream Multi-Stream

Port 0 Port 5 Closed Loop Closed

Dual Bulter Matrix

AC1 1V (A) 1 Column with vertical polarization (V-Pol)

AC7 ULA – 2V (B) 2 Closely spaced V-pol columns

AC5, AC7 ULA-4V (C) 4 V-pol columns

AC2-4, DIV 1X (D) 1 Column with dual-slant-45 polarization (X-pol)

AC2-6 CLA-2X (E) 2 Closely spaced X-pol columns (Quad Port)

1 X-pol center column between two closely spaced x-